The AM Forum
March 29, 2024, 10:46:29 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1] 2 3   Go Down
  Print  
Author Topic: HV supply load regulation issues  (Read 28573 times)
0 Members and 1 Guest are viewing this topic.
AMLOVER
Member

Offline Offline

Posts: 114


« on: May 24, 2017, 01:08:20 PM »

Hi to all folks,

The HV supply consists a transformer with sec at 3850vac/2A. Fully bridged, C input with smoothing cap 30uf/6kv (13 x 390uf/450v) and some ceramic .01uf/10kv. No load 5500v.

Problem is :
If I have 0.3A idling current the voltage drops 10% to 4950Vdc and if I have 1.1A key down the voltage drops another 20% to 4000Vdc.

The mains 230v drop to 220V which is 5% decrease and so is obvious that the rest -25% comes from the HV supply issues.
Transformer's secondary has low R around 7 ohms and the wire is more then sufficient for 2A. It also weighs around 60lbs - 25kg

Another strange thing is that during the modulation peaks - speech - the mains ac current in the primary of the transformer moves to lower from 20A to 16A as it also happens to the dc current from 1.1A to 0.9A.

Any ideas about improving the HV supply and avoid that 25%-30% voltage drop in full load before I add more turns in the secondary of the transformer in order to get the desired 5000Vdc at 1.1A?

Adding capacitance seems too simple to be helpful ...

Logged
SM6OID
Member

Offline Offline

Posts: 220



« Reply #1 on: May 24, 2017, 01:25:35 PM »

Hi!
Just to verify facts,  may I suggest that you check transformer secondary voltage.
Load vs no load voltage.
Logged

RADIO: 51J-4, R-390A, SP-600 JX-21, BRT-400, Set No 19, T-47/ART-13, RF-590, SRT CR91, BC-312D, BC-348Q, HF-8020/8030/8010A/8090,  and much more...

ENGINE: Zvezda M50 F6L (V12), Rolls-Royce Meteor mk4B/2 (V12), Rolls-Royce B80 (inline 8 ) and much more
AMLOVER
Member

Offline Offline

Posts: 114


« Reply #2 on: May 24, 2017, 01:43:17 PM »

SM6OID,

I can check it only by vdc panel meter. I can only see 5500v no load, 5000v with 0.3A and 4000v with 1.1A. I can't check the ac sec voltage with no and full load.
It is difficult to find ac meter for that hv sec (3750vac).
Logged
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2648


Making AM GREAT Again!


« Reply #3 on: May 24, 2017, 02:09:25 PM »

Make a voltage divider out of a series of metal film resistors.  That will allow you to see secondary voltage.  Or, spend 30 to 50 bucks on eBay and get an HV probe.

Also, measure primary V Drop.  My last 3cx3000 project was dropping 10 pct on the primary.  The secondary had the same amount of V drop. Because of that, not much could be done on the amp side to make sure the plate and fil voltage didn't sag, except increase the current capability on the input side.

If your primary isn't sagging 20 pct under load, then look at the resistance of the secondary.  If it's high, not a lot can be done to fix that, either.  More filter C, or CLC filter maybe.....?

Need a bit more info.

--Shane
KD6VXI
Logged
AMLOVER
Member

Offline Offline

Posts: 114


« Reply #4 on: May 24, 2017, 02:45:03 PM »

KD6VXI,

Secondary R is only 7 ohm, primary drops 5% on full load so no big problem.
Secondary voltage dropping - which I will try to measure somehow - or capacitance increase could be a matter of thinking.
I see few topics and find some HV supplies with as much as 150uf of smoothing capacitance but this could mostly influence on ripple percentage and not on supply's load regulation.
I should also mention that the amp's service is as AB1 linear. The voltage's drop however is noticeable before modulation occurs.
Fortunately I have 3 phases mains and filaments plus other supplies are supplied from a different phase in order to be stable.
 
Logged
KA2DZT
Member

Offline Offline

Posts: 2192


« Reply #5 on: May 24, 2017, 03:43:48 PM »

Regulation with cap input filtering is always poor.

You're not going to see 5KV @ 1amp with that xfmr.

Fred
Logged
AMLOVER
Member

Offline Offline

Posts: 114


« Reply #6 on: May 24, 2017, 05:19:55 PM »

Fred,

I agree with you. I was wondering though if I should firstly invest on capacitance or on adding more ac voltage. I will definitely go first for an extra 1000 vac in order to achieve the 5kv dc and I'll see later if I need more smoothing capacitance. I had few problems in the past with L input hv supplies and so it is out of my plans.

Stefano
Logged
WBear2GCR
Member

Offline Offline

Posts: 4135


Brrrr- it's cold in the shack! Fire up the BIG RIG


WWW
« Reply #7 on: May 24, 2017, 05:20:37 PM »

It's not a great insight, but ur drawing 1500watts from the wall and you are seeing a % drop.
The typical "rule of thumb" for CCS operation is a 10% drop.
That's for the supply as a whole, and also for the iron without rectification/filtering.

I'm a bit concerned about the voltage drop on the AC mains with a 1500w draw.
Logged

_-_- bear WB2GCR                   http://www.bearlabs.com
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #8 on: May 24, 2017, 09:20:44 PM »

HV regulation has always been a challenge for me.  I ran into the same issue where the 240 primary dropped only 5% but the DC HV output dropped  MUCH more and ruined the overall regulation.

You certainly want to get to the bottom of this because a linear amp running class AB/B will have its IMD go to pot as a result. The idle current helps, as well as the carrier on AM to provide a "bleeder current" but if the regulation is really bad, it will simply drop further.

I've found that the solution is to first, use a HV transformer that is over-rated for the job by at least X2. There is nothing wrong using a big MO-FO transformer to supply the filter its needed pulses.

Next, a choke input filter will help regulation a lot. This means planning ahead since you will see about 0.9 X your AC HV.  

Next, think about this: What if you used a near-INFINITE sized filter capacitor? Let's say a tera-farad.  :-)   Once this capacitor was charged, under full load there would be virtually no DC HV drop at all for a LONG time. So, certainly, using as big a filter capacitor value as possible will help your regulation.

I like the suggestion of using a resistive divider or getting a HV probe to get in there and determine exactly where most of the existing drop is now coming from.

But generally, if the 240 VAC mains is at least #6 (and preferably bigger), the trouble is either caused by IR drop in the transformer wire, insufficient core size, too little C, rectifier drop when using hollow state or not using an L-C filter.  Watch the drop across the inductor too.

For my 4KV HV supply, I use a pair of paralleled 5 KVA pole pig transformers, a 200 uF oil-filled capacitor bank, a 30H large choke that is switchable in and out and 6A solid state diodes. It took a long time of dicking around, but my overall HV DC regulation is now down to under 8% from no load to full load.  

Once my rig is keyed and the bleeders, idle current or AM carrier puts on the initial load, there is barely any drop under modulation on both ssb or AM. That is the critical issue: How much does it drop under modulation? How much does the audio cyclic rate affect regulation? The less drop, the cleaner the signal will be. (Better linearity, less compression and less flat topping effects).

T


Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
KA2DZT
Member

Offline Offline

Posts: 2192


« Reply #9 on: May 24, 2017, 09:46:07 PM »

60lbs for a xfmr trying to deliver over 5kva might be a little light.  Could be the iron is lacking.  I have xfmrs that are only 1.5kva that weight more than 60ibs.  If you load down the supply for any length of time, how warm does the core get.

I think you have too little capacitance to maintain good regulation.  Do you have any bleeder current and if, how much??

Only real way to get good regulation is to use a choke input filter and a xfmr that has a much higher secondary voltage.  With choke input you need close to 5800 vac secondary voltage to end up with 5KV DC after filtering.

Fred

Another thing,  3850vac @1amp puts a nearly 20amp current in the primary.  What size wire is in the primary??
Logged
W7TFO
WTF-OVER in 7 land Dennis
Contributing
Member
*
Online Online

Posts: 2521


IN A TRIODE NO ONE CAN HEAR YOUR SCREEN


WWW
« Reply #10 on: May 25, 2017, 09:46:11 AM »

If your iron is found lacking, get a three phase one as you say you have it available.  Much better regulation, lower ripple, higher efficiency.

They are usually cheaper and easier to find than single-phase iron of comparable power.  Look for one from old induction heating equipment.

73DG
Logged

Just pacing the Farady cage...
WD8BIL
Contributing
Member
*
Offline Offline

Posts: 4409


« Reply #11 on: May 25, 2017, 09:55:39 AM »

Not all capacitors are the same. If a cap input supply is what you want choose capacitors with very low ESR. 1 or 2 capacitors with good ESR ratings and sufficient voltage capability is far better then the common 8, 10 or 12 stack.
Logged
AMLOVER
Member

Offline Offline

Posts: 114


« Reply #12 on: May 25, 2017, 03:16:15 PM »

WBear2GCR,

In the case of idling current of 0.3A and 1500w consumption I get 5% drop in the mains which is acceptable but I get a total of 10% drop in the dc output of the supply. This is what I should cure...

Tom,

It's true that supplies must be twice stronger then our needs. I think that the super cap, 200-300uf, would be helpful in the modulation peaks like it does in the B class audio amps that I also use in the high level mod transmitter. Things become of course worse with modulation peaks and instead of going a bit positive the dc ammeter goes 200-300ma back.

Fred,

60lbs would be enough at least for the idling 300ma but even there there is significant drop. I will add anyway some 50-100uf, I use 13 x 47k/10w resistors one to each cap for discharging and nothing else as bleeders. Choke input needs huge ac voltage and when there is no load the voltage goes very high. it is dangerous as far I always keep the hv on. The pri wire is 2.7mm and the sec is 1.1mm. The core becomes very warm after 15-20min of continuous transmitting with 100% music modulation.

W7TFO,

Many folks here are using a 3 phase transformer with 3 rectifiers and one cap around 10-20uf - successfully. They have almost no hv dropping and very little ripple for such a small cap.
Some others though, needing higher voltages, use 3 phase mains but with 3 individual one phase transformers. They make 3 different supplies, one for each phase, and all the 3 supplies in series. In this case they also suffer from serious voltage dropping. It seems that the 3 phase transformer is the easier way to make a very stable hv supply. I had spare this one and gave it a try but ...

WD8BIL,

I could test your indication before I order the 3 phase transformer.
I will try 2-3 oil caps in series with 30-40uf total capacitance and see the difference. I thought that uf are uf wherever they come from. Is it true that the oil caps are superior of the stacked electrolytics?  I only use the stacked electrolytics because they are newer and possibly less tired...
Logged
KA2DZT
Member

Offline Offline

Posts: 2192


« Reply #13 on: May 25, 2017, 04:27:16 PM »

With no real load (13x47Kohm very little bleed) the caps are going to charge up the the peak AC voltage which is about 5500 volts.  That peak voltage will drop considerable with any amount of real load.  300ma will easily cause a 10% drop from the peak voltage. I'm surprised it is only about 10%.

That xfmr will never make 5KV@1amp with any amount of real world capacitance.  You're expecting an only 10% drop from the no load peak voltage to a full 1amp load voltage (5500V to 5000V) and that is never going to happen with the set-up you're using.

When looking at regulation you have to start with the DC voltage at the minimum continuous load to full load voltage.  Looking at regulation starting with the peak no load voltage is meaningless, especially true with cap input filters.

Fred.
Logged
Opcom
Patrick J. / KD5OEI
Contributing
Member
*
Offline Offline

Posts: 8308



WWW
« Reply #14 on: May 25, 2017, 05:03:10 PM »

3CX3000 can consume 4800V@2A which is 9600W. A 20KVA transformer is 2x the max load. A choke type filter is the easiest way to get really good regulation. The tests on this supply bear out the simulations for regulation. Regulation is about 5% with the resonant choke setup which also has the lowest transients. Adding a large choke and cap after it gives less ripple, slightly more transient, and about the same regulation. Using just a big 30H choke gives the lowest ripple, about 10% regulation, and quiote higher transients.
Note here that the load goes from 50mA to 2A and is 'snapped' on and off with a (indestructible transistor) switch. This would not reflect real use with voice so the transients would in practice be less. The point is that the choke input or resonant filter with an oversized transformer is what you may want if you want good regulation.


* 3-choices.png (43.75 KB, 949x851 - viewed 313 times.)

* power supply.png (65 KB, 959x1094 - viewed 277 times.)

* ripple-load-transients.png (25.97 KB, 959x419 - viewed 269 times.)
Logged

Radio Candelstein - Flagship Station of the NRK Radio Network.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #15 on: May 25, 2017, 05:35:37 PM »


Choke input needs huge AC voltage is when there is no load the voltage goes very high. it is dangerous as far I always keep the hv on.


No worries if you use a step start. Mine uses a T/R keyed 10 ohm, 200 watt resistor that is in the 240 primary for 1/2 second and then shorted out. It does a few things.... first the surge is greatly reduced especially with my 200 uF cap bank and SS rectifiers. Though the choke, when in circuit, cushions this.

Second, the voltage never soars. It starts out at about 3300V for 1/2 second and then rapidly rises to 4KV. It never exceeds my target voltage.

Third, is safety. I just hate hearing that HV hum when it's on. The step start keeps it off 95% of the time, or however much I have the rig unkeyed.

Fourth, it saves on power - eddy currents and bleeder heat.  I don't feel as guilty when the big rig is on and I'm just listening.... :-)

A HV step start is a really nice item to employ in the shack.

T

Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
WD8BIL
Contributing
Member
*
Offline Offline

Posts: 4409


« Reply #16 on: May 25, 2017, 06:32:26 PM »

Series resistance in the capacitors will limit how fast a capacitor charges with each voltage peak. Limiting this will improve your regulation. Some of the 450uf 450VDC caps I've seen coming out of wherever have extremely high series resistances. Stringing them in an 8 or 10 cap string just multiplies this resistance.

Logged
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2648


Making AM GREAT Again!


« Reply #17 on: May 25, 2017, 07:14:36 PM »

I'll second BIL comments on the caps.

I bought a cheapie ESR meter.

Wow.  Like the bird Watt meter, it's a heart breaker!

I would say the caps are 'bargain' electrolytics.  Or you have some REALLY puny wire interconnecting.

Looking back at your numbers, your DC voltage is sagging to within 150 volts of your transformers AC secondary output......?  3850 AC 4000 DC.  That's horrible.  Sure you don't have a crappy diode?

My last 3cx3000 has a 12kva Pete Dahl ccs rated xformer in it.  It will do 10kw out the snout, pep, for 8 minutes (longest song I had).  HV sags from 6kv to about 5500 under modulation.

This is in a high level xmitter?  If so, those voltage drops explain the carrier shift downward.

I'm afraid adding windings to increase the loaded B+ isn't going to help.....  Your still going to have horrible regulation.


Also, another thing to check.  All screws tight, from the wall socket to the plate choke? I2R means you'll have bad regulation if there is a loose connection.  And that regulation will get worse as the load increases (ask me how I know (stupid computer electrolytics with screw terminals!)). 

--Shane
KD6VXI

Logged
KA2DZT
Member

Offline Offline

Posts: 2192


« Reply #18 on: May 25, 2017, 08:05:56 PM »

The voltage drops to 4KV with a 1.1amp load.  That's a lot of current to ask the 450 volt caps to supply between peaks.  At that load I wonder how much ripple is riding on the DC.  He needs to make voltage divider to lower the DC voltage to a lower level to use a DVM.  Measure some level of DC (say 500VDC).  Then switch the meter to AC and measure the level of the AC component.  Or, just use a scope.

At the 1.1amp load if the DC is still clean without any ripple, the problem is not the caps.  But, highly unlikely the DC under load is clean.

Fred
Logged
SM6OID
Member

Offline Offline

Posts: 220



« Reply #19 on: May 26, 2017, 04:54:30 AM »

Hi!
3-phase is the path to walk, I have 400 V 3-phase.
One of my Harris RF-130 has got the 3-phase power supply, voltage drop is low, even if beeing pushed "above limits".

Once I tried primary regulation in a project, approx 1 kW level, worked like a charm. Design suggestion can be found in mid 90's handbook.

And, another thing to consider, with huge capacitors the stored energy will become as impressive as destructive.
Unless, the design can handle hard faults.

Last but not least, make sure that you have proper equipment to measure what needs to be checked.
Be warned, voltage is high and the stored energy can be phenomenal.
Logged

RADIO: 51J-4, R-390A, SP-600 JX-21, BRT-400, Set No 19, T-47/ART-13, RF-590, SRT CR91, BC-312D, BC-348Q, HF-8020/8030/8010A/8090,  and much more...

ENGINE: Zvezda M50 F6L (V12), Rolls-Royce Meteor mk4B/2 (V12), Rolls-Royce B80 (inline 8 ) and much more
AMLOVER
Member

Offline Offline

Posts: 114


« Reply #20 on: May 26, 2017, 06:52:12 AM »

My set-up is the same like in the schematic below except the 2400uf/450v caps.
The results though are different Sad
I had few 'bargain' 390uf/450v and thought like all of you that with less capacitance the possible negative will be a worse ripple % but definitely the real negative is the extra 20% dropping under load.
The schematic comes from w2dtc.com who supplies 3cx3000s and he has a very acceptable 10% drop under full load with the same transformer secondary ac as mine and the same cap input regulation system.
I use 24 x 10A10 diodes (1kv/10A) 6 per leg of a full bridge.
The supply in the schematic uses 150uf instead of mine 30uf. Can't believe that this is the problem of my extra 20% drop.
Even if Opcom's 3rd option seems simple and manageable, choke input needs a new transformer with at least 5500vac plus the 2.5H choke and care on controlling transients even with the resonant filter cap. In such a case the 3 phase transformer with smaller cap and no choke seems preferable.
There is soft start but for my set-up is more convenient to keep HV always on the plate.
The amp is AB1 linear with carrier indications 1.1A plate, 0ma grid and 30ma screen (bleeder's current). On modulation peaks, grid shows 10ua, screen 100ma and plate a horrible 900ma.
I will check the net for "primary regulation" it sounds interesting.

Stefano


* HV supply W2DTC.com.txt (86.22 KB, 800x600 - viewed 353 times.)
Logged
SM6OID
Member

Offline Offline

Posts: 220



« Reply #21 on: May 26, 2017, 07:07:21 AM »

Hi!
Stefano, when I get back home, I will look in the archive.
AT least I can find the original article. My "own" design documentation got lost years ago when I changed employer. Perhaps it was in QST I found the original drawings. DC regulation less than 1% is easily done.
Logged

RADIO: 51J-4, R-390A, SP-600 JX-21, BRT-400, Set No 19, T-47/ART-13, RF-590, SRT CR91, BC-312D, BC-348Q, HF-8020/8030/8010A/8090,  and much more...

ENGINE: Zvezda M50 F6L (V12), Rolls-Royce Meteor mk4B/2 (V12), Rolls-Royce B80 (inline 8 ) and much more
AMLOVER
Member

Offline Offline

Posts: 114


« Reply #22 on: May 26, 2017, 10:27:58 AM »

SM6OID,

It will be appreciated, thank you.

I have attached my set-up and few meter indications.
The strange thing as I now realize is that when the dcA go down with mod peaks the acA in the primary go down, too and in the same time the dcV in the supply's output go a bit up as also happens to the primary acV.
In mod peaks the amperes dc and ac go down meanwhile the dc and ac volts go up...


* schematic.gif (19.67 KB, 1360x511 - viewed 292 times.)
Logged
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2648


Making AM GREAT Again!


« Reply #23 on: May 26, 2017, 10:34:12 AM »

Put your dvm across the bottom capacitor or two.

The caps / bleed / equalizing R will act as a voltage divider and allow your dvm to make real measurements.

Scale readings up by the amount of caps in series.

Do both the AC and DC test described by another post.  Reply back with results.

--Shane
KD6VXI
Logged
KA2DZT
Member

Offline Offline

Posts: 2192


« Reply #24 on: May 26, 2017, 11:16:14 AM »

SM6OID,

It will be appreciated, thank you.

I have attached my set-up and few meter indications.
The strange thing as I now realize is that when the dcA go down with mod peaks the acA in the primary go down, too and in the same time the dcV in the supply's output go a bit up as also happens to the primary acV.
In mod peaks the amperes dc and ac go down meanwhile the dc and ac volts go up...


Move the ground on the cap string to chassis ground,  you have it going through the meter.  Redo your measurements.

Fred
Logged
Pages: [1] 2 3   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.067 seconds with 19 queries.