The AM Forum

I have been having problems with transformer acoustical 60-120 Hz buzz when drawing rated current from my power supplies.  It gets so loud it sounds like a short circuit and gets into the mic.  I rebuilt the whole HV supply and put in different parts. No difference.    The PS regulation is decent and the ripple is minimal - very clean.  Just makes a racket when more than 300 mA is pulled from a transformer rated at 1A.  I am sure the PS is OK. It is very quiet below 300 mA.

After a lot of testing and troubleshooting I realized that the INCOMING AC 240V mains is flat topping on the peaks. It appears to have harmonic distortion or a phase problem with a poor  power factor.  This is the last thing I expected to find.   As shown in the pictures below the first group is a scope shot with most everything in the house turned off. Then If I put a big QRO TX load and the oven on etc., the waveform does not change. It appears to be occurring somewhere down the line (before the pole pig on the street) regardless of what I do in the house. There is a single 15 KVA pole transformer outside feeding just this house.  I tried some 4 uF caps across the AC lines and had no effect.

I am wondering if the squarewave portion is beating up my xfmr iron, making it buzz. It also makes the PS choke sing.    A local friend took a shot of his own AC line in the second group below and his looks  MUCH cleaner. He has no xfmr buzz problems.

I plan to call the power company to see if this is a problem that can be corrected via capacitors or inductors on their end.  I am wondering if they are simply loading the 7 KV line too heavily somewhere down the line.   Or do I need a harmonic filter?  

Does anyone have experience with this and does my AC waveform distortion look within reason?  I am thinking that a pure square wave put thru 60 Hz iron would ring it like a bell. A sinewave with small amounts of saturation  would be some degree less, but still enuff to cause noise problems.  In the meantime I bought some moving blankets to cover the HV PS cabinet as soundproofing. This helps a lot, but is a PIA.

BTW, if you look at your own AC line, be sure that the scope ground goes to the neutral and not one of the hot AC leads. Please post your AC sinewave if you can.  If you have some distortion do you notice any excessive xfmr buzz under full load?

T

These are pics of my own utility 240V AC drop. Notice the peaks are NOT a classic smooth pointed sinewave, rather, beginning to take on harmonic-rich characteristics:

Here's the new HV PS and my friend's AC sinewave.  Notice how much cleaner his is:



PS: 4KV, 133 uF, weighs about 300 pounds.  It powers three different rigs, so is an important piece to have quiet. I finally got the big linear blower to be silent by mounting it outside - and now this... (sigh)

T

Tom, a 15 KVA transformer seems rather small, even if for only one home.  That provides only around 60 amperes at 240 volts.  

In my neighborhood, most homes have 150 to 200 amp service at 240 volts.  In my case, the pole transformer is 75 KVA, shared by three homes.  75 KVA is about 312 amperes at 240 volts; if equally loaded by the three homes, 104 amperes per home.  This allows for much higher peak load without flat-topping, and the peak loads are usually not simultaneous across the three homes.

I realize in normal use you would not reach this much load, but with air conditioners, water pumps, water heaters, etc. it seems 60 amperes is rather PW.

When you run a capacitor input power supply, most of the load on the service is just on the voltage peaks, when the rectified AC peak is just above the voltage charge on the filter capacitor.  Using choke input, with the additional energy storage and resultant filtering provided by the series inductor, it spreads this load period over more of the input waveform. The more inductance, the less peak current for a given average current.  I realize you can reach a point of diminishing returns, and the resistive losses in the inductor add to the problem.  But that is probably the only thing you can do, other than requesting a QRO upgrade to the pole transformer.

You could also offer to reduce the cost of the upgrade by volunteering to handle the recycling of the PW pole transformer now in service.  You might also mention that the flat-topping and power factor issue might be alleviated with an upgrade, thus avoiding a pole transformer crap-out in the future during one of your January old-buzzard shack-heater transmissions!

Tom..  Do you have any chokes filtering your supplies?   Or just Brute force capacitance?...curious as I am starting to figure my line supply as part of the system....I really haven't had to think much about it before...I will try to post some scope shots when I get the setup going..What is your setup for the scope patterns you posted? Are those loaded line patterns ? with the big rig on?
 Gud Luk..Steve

One more thought.  Have you looked at the line waveform at the main breaker panel, or only on the drop into the shack feeding the QRO supply?  You could easily look at it on a different circuit that is not loaded, while your power supply provides the load on the shack drop.

If it is clean at the panel, and flat-topped in the shack, perhaps a larger wire circuit to the shack would be the simplest fix?  And again, it is not about sizing the wire for the average load, but up-sizing the wire to avoid the voltage drop on current peaks.  Standard size wire for the load current plus long distance from the panel to the shack may also contribute to the problem.

For my larger 120 volt loads, for example the Hallicrafters HT-32 linear, which runs at 120 volts, I use a 240 to 120 isolation transformer to cut the line drop in half.  You may be suffering the same issue with your 240 volt load.

Thanks for the replies, Rick and Steve - this is encouraging.

Well, I got rid of the PS chokes cuz they just added to the noise. My thinking is if there is any saturation/flat topping/ harmonic-rich waveforms, all magnetic iron hardware is going to be affected.  It is quieter without any chokes for now.  That is a clue.

BTW, my AC line scope pics are taken with almost everything in the house shut off. I can't imagine the 15 KVA street pole pig being loaded down. And when I put on a full load in the hose, the waveform does not budge. That's why I think it is occurring BEFORE it even arrives to the street pole pig.

Interesting on your pole pig rating comments. I used to have a 5KVA up there until I complained about my arc welding pulling it down. But most of the houses around here still use 5KVA units.

Rick, do you think there is enuff distortion on my waveform to actually cause this problem?  I've never heard about this before and have not come across very specific info on the web. There is a lot of talk about harmonic distortion and power factors and how to correct it, but no one has actually said it will cause power supply iron to start buzzing like this.

Steve, on the scope hookup:  Just connect the AC neutral to the scope probe ground lead and the scope probe to the hot 120V line. (X10 probe)   Minimal load in these pics.

Rick -

The regulation is great. I measured the 120V-0-120V  in the shack and under FULL load each leg sags from 120V to 119V.  It is very good regulation to the shack and from the pole pig. But it screams like crazy when doing it... :-).

Again, with no load on the street pole pig (everything in the house shut off) the sinewave is distorted as pictured. A heavy load does not effect it.


I am fishing for someone to tell me that I have excessive distortion and yes, it will cause loud xfmr buzzing...   ;)

But you gave me a good idea - I will bring the scope into the cellar tomorrow and close off the 200A breaker to the house.. completely dead.  And then look at  the sinewave connected only to the street pole pig.  I haven't actually done that yet.   Thanks.

T

https://www.testandmeasurementtips.com/measure-powerline-quality-oscilloscope/

Tom I have had some chokes that made noise and others that were silent....I have mostly used what was available rather than what was correct ,but in true hammy fashion was always able to make things go...  Just seems like chokes (in the neg or pos line)  ..just make everything smoother..nothing empirical ...just my observatio....n...let me now what you find...I'm interested.

hi Tom .... when I used to work for TVA as an unlicensed nuclear plant operator (glorified valve ape) I used to study the 500kV step up transformers .... 1200 mvar single phase I think ... 3 used at a time and each tranny about 15 ft on a side .... the nameplate data indicated 3% harmonic distortion .... this made for interesting Tesla coil type effects in the switchyard like walking up to grounded panels in the yard and drawing 2" arcs to fingers with no pain ... most of the other workers were glad for me to take the switchyard job evolutions  :D

The idea here is to look at what is coming into your panel WHILE THE POWER SUPPLY IS ON.
Then you will be able to see if the flat-topping is present on the line coming into the house.  If it is NOT present at the panel, but it is present at the input to your power supply in the shack, then the current peaks are causing voltage drop across the drop to your shack, and NOT in the feed from the utility. Then the simple fix is to upsize the feed to the shack, even if it is sufficient for the average current.  The peak current is only for a short duration of the 60 Hz waveform.

An example case in point:  Back in the late 80s, we had  large HP mainframe, running on 3-phase 208 power.  The computer had MANY switching power supplies, consisting of a capacitor input filter right after the bridge rectifier.  All the supplies demanded max current at the crest of the incoming 60Hz waveform.  Now, to make matters worse, each supply in the system was connected to neutral and one phase, 120 volt into each supply.  The load on the neutral was THREE TIMES the current on each of the individual phases.   The feed to the computer was four number 10, plus ground.  We looked at the signal on each phase, and the neutral, referenced to ground, with a distortion analyzer.  The power supplies were, in effect, current starved due to the excessive neutral current.  Later designs of the computer power supply used 208 volt phase-to-phase, and had inductance in series between the rectifier and main filter capacitor bank feeding the switching power supplies, problem resolved.  But the interim fix for the earlier 120 volt supplies demanding excessive peak neutral current was to upsize the neutral feed from number 10 to number 6, for a nominal 30 amp 208 volt three-phase service.   (Then later version did not even have a neutral.)

The point I am trying to make is that the "metered" current on your feed to your shack is probably well within the scope of the wire size, but if most of the current is in short duration pulses, then expect significant voltage drop over the wire sized for average current.  Kinda like the wiring in your RF tank, no?

So measure the waveform at the entrance panel, with the transmitter power supply loaded normally.  If it is clean there but not in the shack, then you either need to reduce the peak demands with power supply design changes, OR increase the peak current carrying capacity of the shack feeders to minimize the voltage drop at peak current, which causes the flat top waveform at the power supply input.

I guess we doubled, or I did not determine whether the frequency was in use before transmitting.   I must be one of those dreaded automatically controlled data station, providing internet service to the fishermen in the north Atlantic.....  my sincerest apologies, Hi Hi!!

If you are seeing the distortion on the incoming waveform, and you have no loads on the service, then there is likely nothing you can do within your premises to resolve the problem.  The power company must determine whether the distortion is on the 7KV feeder, or if it only exists on the output of your pole pig.  If it is on the feeder, they will need to do further research to resolve the issue.  If it is only on the load side of your transformer, perhaps a turn or two is shorted, and the transformer is already running hot.  A replacement pig, perhaps a more well-nourished pig, say 25 or 50 KVA, would be a welcome upgrade.   Do it for the children!

reminds me of the modified sinewave inverter I once had.  the sound of various devices that were plugged into it were not happy sounds.

Update: My sinewave. that's just one 1 phase but no unusual sounds from transformers or odd behavior on devices in the house or shack.

Bad ground at the pig possibly.  Or bad neutral to ground bond at the panel?

Have you measured both phases to ground?  And measure phase to phase (this could be tricky, float the scope).

We had a problem at work where we dropped our neutral.  All kinds of interesting things happened.  PG&E ended up taking responsibility for the problem and replaced a slew of items in the shop.

I could see waveform distortion happening if the ground to neutral bond wasn't working correctly and your getting an 'artificial' neutral brought about by the equipment connected.

--Shane
KD6VXI

Check the current waveform on the incoming power. This will clearly show if or how much you are 'abusing' the mains.

In lieu of a proper current probe like a clamp-on, you can wind about 3 to 5 turns of 600V insulated wire around the insulated power conductor, terminate the coil with 1K Ohm, and connect a scope across the coil. You will get a decent representation of the current waveform, though it won't be calibrated since it's not a proper current transformer. A low Ohm load keeps the voltage across the coil from rising as it might if left unloaded/open/loaded only with a 10meg scope probe.

There are some examples of this kind of check with the voltage and current waveforms superimposed using a 2 channel scope.
http://bunkerofdoom.com/mil/m35/gen/test1/index.html

The third and fourth ones on the top row maybe of most interest for capacitor-input power supply users.

To a retired electrician the use of the word "drop" for anything except the power company wires that, hopefully, drop from the pole to the connecting splices of the Service Entry Cable to your hose makes me dizzy. The term "drop" comes down to us from the land line telegraph era. You know, your great grandparents' email. Every building served by a public power utility has either an aerial service drop or an underground service lateral. In either case that belongs to your power utility. The only reason that is worth knowing for this thread is that when you talk to your Utility's staff you'll want to use basically correct terminology or you will be subjected to that dreaded response "Have you're electrician call the engineer at this number ***-***-****. Have a nice day." It IS important to sound like you know what your talking about. If they are unresponsive or stonewall you, know that there are a number of people they do not want to hear from about your problem because it will make your problem more urgent than anything but several customers without service. Those people are, in increasing order of urgency: Your local government council member, selectman, or whatever that person is called in your community;  your State representative; Public Service Commission, Public Utilities Commission, People's Council; State Senator; and, especially not, a State Delegate or Senator who sits on a committee that deals with public utility issues. That last will make an obstinate utility official come to life from the grave!

The engineers at the state utility regulating agency will understand your scope traces so if you can actually get to talk to one of them and send them your traces that can be quickly curative. These things I learned experientially during 40+ years as an electrician in all kinds of different premises from homes to pharmaceutical laboratories and from College campuses to solar powered communications equipment shelters on ridge tops in Argentina. I would suspect that somewhere upstream of you, on the power distribution phase that eventually supplies your pole pig, there is a manufacturer with an number of variable frequency motor drives, a user with a fairly large number of computer power supplies, or both. Their up-sized neutral may make there equipment love life again but it will not erase the harmonics they are producing off of the distribution network. Once the POCO is forced to face the problem they will often force the offending customer to clean up their power usage by power factor correction. Since the state regulators usually allow the utilities to apply an up-charge to such customers until they correct their power factor an old GI Drill Sargent slogan applies. "That there is called a pit. That there is called a rope. The alligator is called an incentive!"

I have faced a couple of these issues in my time. A good example was for a dairy farmer who's electricity supply made their cows unwilling to drink from their water cups nor the stock tanks. How much milk would you think dehydrated cows would produce? When I had the farmer attach his power take off generator and tractor to his service disconnect so he could continue his previously crippled operations, opened the POCO service drop conductors from the pole pig including the neutral, and showed the utility regulator's engineer that the shock voltages were still present and that ~45 amperes was flowing on the briefly reconnected neutral conductor with the tractor powered generator not running; he found that totally convincing. The POCO had to supply a generator at their expense until they cleared the high resistance connections on their neutral all the way back to their sub station. The were also forced to convert to high impedance grounding of their distribution neutral for a quarter mile each way on the road that served that farm. Then had to change the pole pig to a fully isolated type that allowed the creation of a truly derived system were the only connection to the POCO's distribution was magnetic.

As soon as the PTO generator was supplying the farm and the neutral was open the farm couple started literally crying seeing that their animals were no longer being tortured, milk production went up by 50%, and they treated me as if I were the angel Gabriel. Once the POCO had installed the new pole pig and they saw that the change was permanent they gave me a $600 bonus. Since that was 1974 that would be over $3000 today. The farmers got me a lot of customers just by bragging about the production increase at the feed store and the dairy coop. After the third farm the regulatory engineer and I were on a first name basis. By the fourth farm he would take my call and just call the POCO engineer with the two words I needed him to say. "Fix It!" Some of those farms had internal problems caused by grounding the neutral at multiple buildings which had no Equipment Grounding Conductor (EGC) in the feeder. That was how that was done then. Adding an EGC to the wiring from the Yard pole; on multi building premises the conductors from the service at the customer owned yard pole onward belong to the property owner: to each buildings disconnecting means and bond only the EGC to the buildings grounding electrode array solved those problems.

The "take away" is that power quality problems can be truly convoluted in their causes and their remedies. Do your home work, load for Bear, and be ready to drop the hammer to get what you need from the electric utility. Find the engineering or political pressure point and use it!

--
Tom Horne W3TDH

Great info!

Pat, I read your link and will make up a current sampler using the coil.

Tom, FB on the way to approach the utility company.   I'll do some more reading and gets some pics of the current waveform.  Yes, must be loaded for bear ASSUMING the harmonic contamination is the real problem. 

Bob:  Your sinewave picture looks similar to mine! Mine is slightly worse in saturation on the peaks, but I am concerned you are not hearing any buzz.  What happens when you load the 813 rig up to max and run a tone thru?  Do you hear ANY increase in buzz from the plate xfmr or choke or does it stay well behaved?     The main issue here is:  am I barking up the wrong tree thinking that the small amount of AC harmonic distortion is the cause of the loud buzz?    I have replaced the HV PS transformer, rectifiers, chokes, filter caps, etc., and see the same results.  The buzz must also be wasting power with what sounds like horrendous eddy currents and general resonant 60-120 Hz noise.

In the meantime, a few moving blankets on the cabinet helps to stifle the noise for now.

I will make the current sample test later today and see what it looks like.

Thanks for all the help.

T

Shane:

What should I look for to identify a bad neutral or pole pig ground or by "measuring both phases to ground?"    Could you elaborate?

I am getting a slight voltage sag from about 120V to 119V under load on each leg, about identical, which is pretty good in the shack.  I checked the ground, neutral and hot connections and look FB so far.

IE, what would a VOM or scope show if the neutral or grounds had problems?


T

Tom,

That's kind of a hard one, but this may work for you:

Turn off all your breakers.  Including the service main for safety.

Measure from a 120 plug the resistance from the neutral pin to ground.  Should be under a few ohms.  If not, then you have a bad bond somewhere, or a mile of wire between the service panel and the outlet.

Go to a known opposite phase plug and do the same thing.

Check your sinewave from phase to phase.  If the sinewave is clean phase to phase, chances are pretty good that you have a problem phase to neutral somewhere.  I've seen that happen when the service triplex / drop / whatever one wants to call it (line from transformer to panel) has corroded connections.  You are then pulling neutral from the transformer through the lossy earth.  And anything else plugged in to the outlets.

However, since your phase to neutral is pretty consistent, it's probably not "faking" a neutral from things in the house connected line to neutral.


I've had more than one rental house that I've had to go back and tighten every connection up.  Since the triplex usually uses an aluminum wire for the neutral (with a couple strands of steel for strength....  I've killed a good pair of cable cutters before I found this out :(  ) it's easily corroded. 


--Shane
KD6VXI

Hi Tom,
Yes, I was surprised to see that waveform too.  And just to make sure I was not having some weird grounding thing with the scope, I took a filament transformer and connected it to the mains and measured the secondary output. Nothing changed. Had the exact same waveform only smaller in amplitude.

I haven't tried the 813 rig yet but I have loaded up the Ft102 and the Viking2 and both transformers do buzz but they have always buzzed under load and I never found it annoying. In fact, both would be inaudible on air that's how quiet they are.  I'll fire up the 813 rig and see what it does.  I've never noticed it because of fan noise in the rack.  

Tom, I was going to say measure for AC voltage between neutral and ground but I don't know what that would reveal.  But I did that myself and I measure 0.21VAC between them.  What that means?  Damned if I know but in theory one would think they should be zero?? is that good or bad? dunno and that is about 15ft from the panel.

If you used a DMM then the high impedance will measure what we in the industry (at least here in my local hall) call ghost voltage.

It's voltage inductively coupled from lead to lead.

That's nothing to worry about.  However, a low impedance meter would be better.

--Shane
KD6VXI

Shane said:Like a Simpson 260!

I have both a Simpson 260 a good friend gave me and a Radio Shack I inherited from my Grandfather.  I tend to use the Radio Shack due to sentimental mental reasons, but it's a high impedance one too so I end up back with the Simpson.

--Shane
KD6VXI

Still remember that nice line of Simpson 260's in the tool crib in shop class, darn near 50 years ago. Can you say timeless?

John

I'm still looking at other possibilities as to why the xfmr is loudly humming.

Here's some test results. Maybe someone will figure out what is happening.  Bear in mind that I substituted or replaced all of the PS components and cannot find anything wrong with its performance except for the loud buzzing that emanates from anything with an iron core, be it transformer or choke.

Tests:

1) I operated the xfmr alone with 240V with a 5K resistor across the output (4KV @ 800 mA)  - quiet.   Also tried 2.5K load (1.6A @ 4KV)  This is over the xfmr rating.  QUIET.

2) Attached the fullwave diode stack with 5K load from diode output to ground (800 mA load)   -  quiet

3) Attached four different filter caps to the diode stack, trying some alone at 30 uF -   quiet UNTIL more than 300 mA is drawn. At 800 mA, LOUD!  The HV PS does NOT like a filter cap with more than 300 mA drawn.  The xfmr is rated at 1A.

It is not laminations cuz I tried two different pole pigs and now using a 3.5KV @ 1A transformer.   The diode stack has been changed as well as the filter caps.   The ripple is normal under full load as well as good regulation from 4900V to 4400.  (full load)  BUT, the xfmr (or pole pigs) make a very loud 60-120HZ harmonic acoustical sound that sounds like a dead short when heavy current is pulled.

The strange thing is if I could run this PS without a filter capacitor (rectified pulsing DC) it would run quiet. But add the filter cap and all hell breaks loose once a load over 300 mA is pulled. It is a linear increase in noise. At 800 mA it can be heard in the other room.

I tried another power supply today powering my 3-500Z amp. I could not detect much acoustical hum at all with a 500mA load, near its rated limit.  So my idea about AC line harmonic distortion seems suspect since this other supply seems to handle it well.

Schematic below. As simple as it gets.  Ripple clean and regulation good.  Primary 120-0-120V sags only one volt for each leg under full load. Perfectly balanced. Nice AC regulation, but LOUD as heck.

BTW, I tried a jumper between the primary CT and the secondary CT. No difference in hum or performance.

This is crazy. Any ideas?

T

** Heck, the way I'm going I might be sticking my head into one of these ...  ;D

https://www.ebay.com/itm/New-PSIB27-Sound-Recording-Booth-Box-Studio-Soundproofing-Foam-Shield-Isolation/362969674891?_trkparms=ispr%3D1&hash=item5482adb48b:g:PtkAAOSwwJ5d5sde&amdata=enc%3AAQAFAAACgBaobrjLl8XobRIiIML1V4Imu%252Fn%252BzU5L90Z278x5ickkBSh1VzQSTzkTiSV5EE%252FHQQwCYEnYBC%252BX%252F5zLsufAcGX9OCceE1xua36PJKnIJEXZ1An%252FtwceJlSjZFQ9etiPUsVELaCyoEYQpSI4thHVhdst8R9Gma2Zytxh0gpi0rAy1XHwXGNMcuPgb3Xg3Xxm3Bv9rrJM%252BsCU%252Bu3IXMAGZiw5wld3ifLVzLnqQjxsGItNYtgOo7gZuNAghnIxNrczg12P29u1Si6OpWtdU8g2LY6STYXKvs%252Bt34ppcnJI8RhCiuspvGCOa2lgoLp8RcWQVdCs%252Fy0%252B6G6rAW6WkZ%252BqIKi%252F0DTn4ukn7KnVtsFz7HudIahhB80WpnNQ5ESJnymgheWKGf6qS5e6twErfv6SNb%252Fr6DUvaoJlIMBBTi0sTHukQTo1SthmQWW67gircvhS5N2uttaXgw4piBJgK0Z8URA1w8Cn12b6DeBGy43L44aLIIZvC58DnTwZRv%252BpvKaB0bAAXo%252FfHYNvAnr4ZGVtptLxHOGrzALC1Lu4yWpWB9ICfxX0V1x%252Bk0Bb2qE32hdTbE1faS0w0p3sr8%252BSt7tKfB4o0rLGPa9pTgHgA4Xbf%252Bde7828QEUCebWgAroFSZLCIXVTRoNlzPOoDQwvXpTLgIZ28r0gj8smgkz9%252BPrUvOI92OO%252Fwy7LbeBpLr60HcRxbr6Ht8cv4RWEC1I5GJA0sv%252FHM26SNLkMJXDc3iVIiOYj4t0GjJ8leHjhLfFaxvu3HN44FpOc72wl54ZkHMD5407XdTY8QglntR6k4CUP2AKPl11DOtckFy3ESXO1AgEjvAtG1j4xVsuQX9SIHVUkbDA%253D%7Ccksum%3A362969674891f7aee62e158440568653625de3e98d91%7Campid%3APL_CLK%7Cclp%3A2334524 (https://www.ebay.com/itm/New-PSIB27-Sound-Recording-Booth-Box-Studio-Soundproofing-Foam-Shield-Isolation/362969674891?_trkparms=ispr%3D1&hash=item5482adb48b:g:PtkAAOSwwJ5d5sde&amdata=enc%3AAQAFAAACgBaobrjLl8XobRIiIML1V4Imu%252Fn%252BzU5L90Z278x5ickkBSh1VzQSTzkTiSV5EE%252FHQQwCYEnYBC%252BX%252F5zLsufAcGX9OCceE1xua36PJKnIJEXZ1An%252FtwceJlSjZFQ9etiPUsVELaCyoEYQpSI4thHVhdst8R9Gma2Zytxh0gpi0rAy1XHwXGNMcuPgb3Xg3Xxm3Bv9rrJM%252BsCU%252Bu3IXMAGZiw5wld3ifLVzLnqQjxsGItNYtgOo7gZuNAghnIxNrczg12P29u1Si6OpWtdU8g2LY6STYXKvs%252Bt34ppcnJI8RhCiuspvGCOa2lgoLp8RcWQVdCs%252Fy0%252B6G6rAW6WkZ%252BqIKi%252F0DTn4ukn7KnVtsFz7HudIahhB80WpnNQ5ESJnymgheWKGf6qS5e6twErfv6SNb%252Fr6DUvaoJlIMBBTi0sTHukQTo1SthmQWW67gircvhS5N2uttaXgw4piBJgK0Z8URA1w8Cn12b6DeBGy43L44aLIIZvC58DnTwZRv%252BpvKaB0bAAXo%252FfHYNvAnr4ZGVtptLxHOGrzALC1Lu4yWpWB9ICfxX0V1x%252Bk0Bb2qE32hdTbE1faS0w0p3sr8%252BSt7tKfB4o0rLGPa9pTgHgA4Xbf%252Bde7828QEUCebWgAroFSZLCIXVTRoNlzPOoDQwvXpTLgIZ28r0gj8smgkz9%252BPrUvOI92OO%252Fwy7LbeBpLr60HcRxbr6Ht8cv4RWEC1I5GJA0sv%252FHM26SNLkMJXDc3iVIiOYj4t0GjJ8leHjhLfFaxvu3HN44FpOc72wl54ZkHMD5407XdTY8QglntR6k4CUP2AKPl11DOtckFy3ESXO1AgEjvAtG1j4xVsuQX9SIHVUkbDA%253D%7Ccksum%3A362969674891f7aee62e158440568653625de3e98d91%7Campid%3APL_CLK%7Cclp%3A2334524)

One of these might help peek into the dark corners of the mystery:

Do you connect the primary CT to neutral? What happens if you disconnect it and just go line to line across the 240V primary?

Are you sure all these transformers are 240V rated and not 208V rated?

Have you tried inserting a 50 Ohm resistor between the cap bank and the rectifier output to somewhat limit peak currents and does that change anything?

Can you use a big 240V variac to bring up the mains voltage from zero to full into the primary of the unloaded transformer and see when it starts buzzing? (secondary not connected and also the pri CT not connected)

Hi Pat -

Those are good suggestions..

These are actually two identical transformers, 3500 @500 mA each. The primaries are in series and the secondaries in series.
Each has a tag showing a 115VAC primary.  I am running 119V under load on each primary and the balance between the two is nearly perfect.

The pole pigs were 240V pri each. I had only 120V on each one and it still buzzed under heavy current loads.

The pri CT is presently connected to neutral. I have tried the 240 across both without the CT and the buzz  noise was the same.  I also tried a large Variac rated at 50A, 120 VAC across one leg and the Variac started to buzz too, just like the xfmr and choke when I used them.  Any magnetic core buzzes in this supply.  I figured the pri CT would help the balance, but it buzzes whether the CT is connected to the neutral or not.

I will try the 50 ohm resistor between the diodes and the cap. I tried it with 10 ohms before but will increase it to 50 ohms to see what happens.


You said: "Can you use a big 240V variac to bring up the mains voltage from zero to full into the primary of the unloaded transformer and see when it starts buzzing? (secondary not connected and also the pri CT not connected)"

Reread my last post about this.  The xfmr stays quiet with a full pri and sec voltage and a full load across the diode stack. It only acoustically buzzes excessively when the last component, a filter cap, is added and the current draw is above 300 mA.  Otherwise the xfmr is near silent.  When everything is connected and working as shown in the schematic, it remains near quiet when pulling a small 100 mA draw.  The 240K bleeder pulls only 17 mA at 66 watts.

T

120V 60 Cycle  looks lumpy...

Edit:  My neighbor has a grid tie solar system........

Do any of your neighbors have a grid-tie solar power system?  Those use a switching topology to achieve the 60Hz, perhaps that is where you are getting the distortion, even across multiple distribution transformers. ???

73DG

interesting on the grid tie solar system.  I have 2 on either side of me.

A cap input filtered power supply draws current only during the peak of the sine wave. It essentially clips it. Your 800mA current draw looks like the attached image to the diode. Look at the graph attached, from Duncan's PSU designer, a free tool on the Duncan amps page. The current is almost 3 amps, and it is for a small part of the time. This is with a generic HV diode type, there's no stacked up set of standard diodes available to simulate.
Point being, your peak current is a sharp peak, not a sine wave, and is much more than 800mA. So it makes the iron sing. If you put in a choke, what then? Or maybe much less capacitor, and a choke input? Check the second attachment...

Ed


PS, before you say anything about the shape of the waveform in #2, it is the current in ONE diode. The current in the inductor is a sine wave, the other half of which is provided by the other diode.

Guys,

Thanks again for the suggestions.

I spent some more time trying different chokes, cap input, choke input and found that a choke simply made the regulation worse due to IR drop. It did not help the buzzing much and brought my HV down below 4 KV to <3500, etc.    ( I need 4KV at minimum.)     A 50 ohm resistor from diodes to capacitor made little difference in noise - just poorer regulation.

So bottom line is I'm using a cap input supply like the schematic. The ripple and regulation is excellent. I built an insulated box around the two transformers inside the cabinet.  I sat the transformers on a 2" think piece of Styrofoam. I also used 3"  foam rubber around the sides and top of the transformers.  I ran some real load tests using the rigs and found the acoustical buzzing is now down to a quiet level. It is there but almost in the ambient room noise.

I ordered a bunch of 6' X 8' moving blankets and will cover the outside of the cabinet neatly with them. I'm expecting the noise to be close to non-existent once all sound paths are gone.  The insulation techniques were my last choice, but it will work FB.  

Also plan to look into the utility company harmonic distortion area. That will take time - but if it helps further, maybe I can get rid of the blankets.  Dennis and Steve - OK on the grid-tie solar system potential harmonic issues. I will check that out.

T

Tom...Was your power supply quiet when it was first in operation?  Your sinewave scope  pic looks like mine and the others.....I just built a powersupply capable of 9kv, I've run it as far as 7 KV with my lumpy sine wave input..it is a brute force cap input on a variac...no noise problems...I know you will figger it out...You may become an acoustical expert in the meantime...Good Luck  73

Tom,
I had a chance to load up the 813 rig  last night.  I turned off the fans and put the back cover back in place and although I hear a buzz, I have to be right on top of the rig to hear it and from a few feet away it's hardly noticeable at all. It's quiet for the most part and would not be picked up by a mic.

I'm wondering if your cabinet is resonating like that of a guitar string on an acoustic guitar at 60/120hz.  A vibration  table would be a nice thing to try here.  Probably a waste of time but it might be worth a try to put a loudspeaker in the cabinet and run some tones through it say from 20hz to 200hz and see if there is an audible peak especially around 60/120hz area.

Am I confused or did you say you have two transformers and not one big transformer? If its two transformers and they are next to each other wont there be some weird things happening with the fields in close proximity? The fields will increase in relationship to the load.
Would speculate that although you have identical transformers maybe under heavy load weird things may happen that won’t happen with one transformer. If you are running just one transformer do you still get the noise under moderate load?

Steve and guys,

No, there was always buzz - when using the original used 10 KVA pole pig transformer, I had hum buzz back in 2010. I built an insulated cabinet around it. That kinda fixed it for 10 years. But last month I tore it down cuz I didn't need such high voltage or desire heavy weight anymore and replaced it with two smaller 3500V 500 mA (rated for each one) transformers. It buzzed again.  It buzzes no matter if I use one or two transformers. I tried tightening the laminations, NG.  I thought they were too close together causing magnetic anomalies, but moving them far apart shows no buzz improvement at all.

Bob - OK on the 813 rig being quiet. And, yes, the speaker cabinet is an issue. The xfmrs are sitting on a 1" piece of wood and 2" of styro, so did the best I can.

As for the original pole pig - it shud be quiet when sitting in a big tub of oil, especially when it was overkill at 10 KVA. But it buzzed terribly, and barely usable with the insulated enclosure.

So, bottom line is the newer, smaller transformers are fully packed in 2-3" styro or foam insulation.  I tried the 4X1 plate modulated rig and my big linear - and yes, the buzz level is acceptable now. Just a hint of buzz. This is a band-aid approach, but believe me, I tried everything. New parts, substituted parts, bridge rectifiers, fullwave rectifiers, different grounding, checked all connections, different caps, tried several big chokes, etc.

The bottom line is now I have virtually no ripple with 133 uF cap input and the regulation is within specs, from about 4900 to 4400V under normal load. Under excessive load it will drop to 4200V, but it's not needed.  The acoustical insulation solved the immediate problem until I improve it further by adding the insulated moving blankets when they arrive and/or call the utility to see what they can tell me about the harmonic sinewave.   The sinewave distortion isn't too much different than what others are seeing. But I still do not know the answer to the question:  Can SOME harmonic distortion like I am experiencing be enuff to cause acoustical hum/buzzing under heavy loads? I still don't know.  IE, if I started with a perfect sinewave, would the buzzing increase substantially as the waveform was changed to show some harmonic distortion on the peaks?  I still dunno.

But the bottom line is that the PS is now quiet enuff to move on and can only get better with the blankets or utility efforts.  However, for now, the REAL solution still remains as an unsolved mystery - and that kinda yellowfies me. ;)

T

I'm glad you have it working acceptably. Your power sine wave is really not very bad but there should be standards. I was surprised that it caused buzzing on three different transformers, but who knows. The only time I have had serious buzzing was on the older square wave inverters. 'modified sine wave' inverters are better but still buzz. On inverters that make trapezoid waveforms, which are only a little worse than your utility, the buzzing was minor. This was all 100-500W stuff so nothing like your levels. Real mystery!

P.S. can you share pictures of the transformers involved, the buzzers and non buzzers?

I tried yet another method to help reduce the transformer buzz as suggested by Frank / GFZ with great results:  

There are two HV transformer primaries in series and the secondaries in parallel to get the proper 4KV voltage and current required. When running the secondaries  in parallel, Frank thinks the circulating currents may be adding to the buzz problems. I put together two separate bridges -  one for each winding and tied the bridges together in parallel so the windings were isolated from each other.   I used (120)  6A , 1KV diodes; 15 in each leg of eight legs.

Fired up the 4-1000A plate modulated rig at full strap and gave the supply a close listen.  Lo and behold the transformer buzz was almost gone. I could hear a faint buzz - about the same level as a transformer running with no load.  Now the only buzz I hear is the breaker panel and the step start relays buzz, which is way down. IE, the HV transfomers are quiet enuff to get swamped by other things in the shack.  The RF carrier on the scope shows very little acoustical hum which I can probably cure with some foam around the breakers and step start.

So, if you ever come across this buzz problem when using two transformers, this simple circuit may help.


I think there are three reasons for less buzz:  

#1: There is no secondary circulating current. Frank says that no matter how closely matched the transformers are, there is always going to be imperfections enuff to make them different enuff to generate secondary circulating currents when wired in parallel.

#2:  With bridge rectifiers vs: fullwave, the transformer works on both sides at once, sharing the load.

#3:  With the secondaries  in parallel, each one runs at 1/2 the current for a given power output, thus less buzz.

Less peak power drawn - more average to keep transformers working together pulling the load, thus less noise.


It all adds up to quieter operation despite the slight AC sinewave harmonic distortion I have to deal with. I may still call the utility to see what they can do to improve it - now that I can live with this.

** One last thing:  The regulation is also better, which is very important to me.   I have gained about 300V less sag under full load. I have another 55 uF @ 5KV cap I might add making it a cap input filter at 188 uF total, which is a lot of storage capacity.  Can someone say "Jumbo Joules?"


T

That is excellent Tom....but why  do the breakers and relays still buzz?

Hi Steve,

The big relays (actually contactors) and AC breakers buzz partially cuz they are all running on AC in the coils and there is still some AC harmonic distortion and acoustic resonance generated by the contacts.   The breakers - I can't do much about except move them away (not)  or cover them with foam rubber.  But the step-start relays; I plan to add a diode, resistor  and capacitor to make the AC coil voltage pulsating DC and see how much that helps. I already have the step relays in a foam lined box. They are also generating more noise when the HV current increases, showing that the magnet sensors are creating bigger fields.

Essentially at this point I have no more hum and buzz that any other QRO shack - just that I am trying to achieve a quiet room with some audio processing.

T

If I were at work, and this kind of problem was brought to my attention, I'd install a 'Dranetz Power Quality' recording analyzer at the utility entrance, and let it run.  Do you have any friends in the Instrument / Electrical shops of factories?  If so, they probably have a power quality recording meter.  They will show waveform (as you show), but also display stuff like THD of current & voltage.  They come with CT's & PT's (current transformers you wrap around your phase conductors, wearing your 8 cal PPE, and voltage sensors).  They usually record a value every few seconds over the period of a week or so, and you can know the worst case distortion, AND /OR the distortion timing.  If the noise is centered around 360Hz, it's (as was suggested above) some carrier noise from a VFD on a variable speed motor.  If the distortion rises and shrinks with the sun, it's either a solar array system, and/or it's someone's small business connected on the circuit common to your home.

A suitably size 'line reactor' applied is very helpful for this kind of problem.

Fluke 738 Power Quality Analyzer  / Dranetz DPBG106.   Be very careful using an O-Scope on utilities measurements.  Most of them aren't CAT III certified (which means if you inadvertently touch something, or have an insulation failure, the meter can blow up in your hand). 

I'll offer one more idea:  I don't know where you live, but most utilities have a power quality troubleshoot organization who will help with these issues for a modest fee.  DES (Duke Power Engineering Services) is the one I can remember off the top of my head.

 

HV secondaries parallelled  :o