DX-60 hum

[n1ps]

I am pretty sure there are some pros on this forum who know this transmitter.  I picked this unit up about a year ago and got to it in the last few weeks.  It hummmms away.  Having worked on many transmitters in the past, the hum did not concern as it is the usual suspects in the power supply or around the heaters in the speech amps.

I noticed both 60 cycle hum and PS ripple, with both being considerable.  First up, the PS,  Replaced all of the power supply except the xformer.  All diodes, caps, etc were replaced.  Added some additional capacitance (5X that of the prints) and even replaced R34 with a choke.  Improved ripple but still there and noticeable as a buzz in a monitoring receiver. But the buzz is at a fairly low level compared to the bigger 60 cycle hum.

The 60 cycle component is far worse.  Tens of 60 cycle volts coming out of the 1st tube as seen on a scope.  Again attacked using the usual methods, checked connections, tightened grounds, rerouted the filament wiring and ran heater grounds back to the transformer.  No affect.  Tried DC filaments with no affect.  Functionally the TX works well.  Good audio and output power after some mods typical for what is published on the web.  It seems there must be some AC currents running around the chassis.  Also tried grounding the chassis with a heavy braid. 

Now, this is my first encounter with a voltage doubler power supply.  I do understand a VD can be tricky.  The only ground on the transformer is at the filament winding, which I resoldered and tightened the connection to the chassis.  Knowing the PS is a compromise with the doubler etc, I know these things can be made to work well.   Just not this one at the moment.  Also, power consumption is about right. 

Out of ideas for now.  Looking for new ones.

TNX
Peter

[KA2DZT]

What happens when you pull the first audio tube?  The hum should stop, if it doesn't then I'm not sure what it could be, it's been 50 years since I last worked on a DX-60.

Someone here will have more answers.

Fred

[n1ps]

yes Fred...setting the gain pot to zero kills the 60 cycle hum
tnx
p

[KA2DZT]

I took a look at the schematic, 12AX7 first two stages of the audio.  Pretty much a standard few stages,  check the decoupling cap on the B+ line between the two stages.  I think it's C-33, if it's open you may be getting motor boating which can sound like hum.

Fred

Took another look,  try shorting the mike connector and also what happens if you pull the 6DE7 or if you disconnect the feedback line from the 6DE7.

[n1ps]

yes all replaced and beefed up.  Much of the speech amps has been rebuilt with new.

[KA2DZT]

yes all replaced and beefed up.  Much of the speech amps has been rebuilt with new.

Running out of ideas.  It's rare, but don't overlook that a new part (caps) could be defective.

[N2DTS]

Have you changed the tube?

You know, you can not have a rig that goes down to 20 cycles and have it hum free without going crazy with keeping it out.
Tube will pick up hum through microphonics, the high impedance d104 type input will pick up AC from miles away.
Did you try a resistor in the cathode of the 1st stage? Sat 1000 ohms?
Run the filaments off DC and rubber mount the tube socket.
Shield all the first two stages wires and shield the tube.

You would have to look at what they did in the old tube hifi preamps to keep the hum out.

When I did up my 32V, I had to run the 1st stage filaments off DC.

Most ham rigs did not go much below 300 Hz stock.

[KA2DZT]

Have you changed the tube?

You know, you can not have a rig that goes down to 20 cycles and have it hum free without going crazy with keeping it out.
Tube will pick up hum through microphonics, the high impedance d104 type input will pick up AC from miles away.
Did you try a resistor in the cathode of the 1st stage? Sat 1000 ohms?
Run the filaments off DC and rubber mount the tube socket.
Shield all the first two stages wires and shield the tube.

You would have to look at what they did in the old tube hifi preamps to keep the hum out.

When I did up my 32V, I had to run the 1st stage filaments off DC.

Most ham rigs did not go much below 300 Hz stock.

All good points, he did try DC on the filaments.  In my HB rig I run the first audio tube filament off a 21 volt DC supply.  The hum reduction is better when the filament is run on a higher DC voltage with respect to ground.  The filament runs from 21 volts to 15 volts.  I also use a 12AY7 which is a better tube for noise and hum.

Adding a cathode resistor may help,  mine is bypassed.  Another thing is he may have to use the one ground point technique.  Running all the audio grounds on a heavy buzz and then grounding it at the mike connector.  All this may be necessary especially if he widen the audio low pass.

Fred

[N2DTS]

The 32V3 had one side of the filament  grounded, and to get DC I just put a diode and cap at the tube socket, picking a cap value that gave me 6 volts at the tube socket with the tube in.
Otherwise, all the low level audio circuits were in a little box and the tubes had shields on them, with no power supply components in the area.
That had the hum levels quite low.
The DX60 is a small rig and it may be hard to keep hum out of the audio.
Does the transmitter have audible hum (local, not over the air)?
How far away is the power transformer from the 12AX7?
Is it shielded?

If you turn the mic gain down or pull the 12AX7 and the hum goes away, its not in the power supply.
What happens if you ground the mic input?

[W3GMS]

Hum in the audio on DX-60's is somewhat of a common problem.  The problem basically is due to ground loops and long audio lines picking up hum fields.  Getting one side of the filaments off the chassis is the first step.  As mentioned stay coupling into the high impedance audio lines is another part of the problem. 

Rather than me trying to restate all what Tim, WA1HLR did to get rid of the problems it would be better to contact him.  I know part of his anti Hum mod relocates the gain pot.   Its been awhile since I heard his dissertation on the subject so its better to contact him. 

Some unmodified DX-60's do sound better than others from a hum perspective.  If you extend the low end response the problem obviously becomes much more noticeable.

Joe, W3GMS 

[n1ps]

TNX  all for the comments.  I think the underlying issue is the power supply.  I need to reduce the ripple further.  The only way is to find a TX type choke of the right value, which I dont have at the moment.  I tried a RX type, but it got some coupling from the xformer...contributing  to the 60 hz component. 

p

[AB2EZ]

yes Fred...setting the gain pot to zero kills the 60 cycle hum
tnx
p

Since the rectifier is full wave, any hum caused by residual ripple on the output of the DC supply should be 120 Hz (and harmonics of 120Hz). [You might want to check to see if both halves of the rectifier tube are working].

Are you sure that the hum is 60 Hz hum? If it is 60Hz hum, then it is not being produced by ripple on the output of the power supply.  It is being caused by:

a) The 60Hz AC on the filament of the 1st preamplifier stage

or

b) Magnetic coupling from one of the transformers into the wiring from the microphone connector to the grid of the first preamplifier

or

c) The external microphone cable is picking up hum from local 60Hz magnetic fields

or

d) If the microphone output is passing through any external audio equipment before going into the DX60's microphone connector, there may be 60Hz hum produced by "ground-loop" problems. In that case, you could fix this by using an inexpensive audio transformer to isolate the external audio equipment's output from the audio input of the DX-60

 
If setting the gain pot to zero "kills" the 60Hz hum, it would be helpful to know (as others have pointed out) if disconnecting the microphone from the microphone input connector of the DX-60 also kills the hum (with the gain pot not set to zero). If disconnecting the microphone kills the hum, that would indicate that the hum is coming from the microphone or the audio processing equipment (if any) between the microphone and the microphone input connector of the DX-60.

If disconnecting the microphone from the microphone input connector of the DX-60 does not kill the hum... then try disconnecting the cable that runs from the microphone connector to the grid of the first preamplifier stage (disconnecting it at the tube socket). This would help in determining whether the hum is being introduced into that cable, or whether the 60Hz hum is being caused by the filament having 60Hz current running through it.

Stu

[W3GMS]

TNX  all for the comments.  I think the underlying issue is the power supply.  I need to reduce the ripple further.  The only way is to find a TX type choke of the right value, which I dont have at the moment.  I tried a RX type, but it got some coupling from the xformer...contributing  to the 60 hz component. 

p

I went through the power supply in my DX-60 in relationship to solving the hum and it was not the problem.  When you do the changers per Tim's instructions, your hum will be gone and nothing needs to be done in the power supply area.    Now if you have bad filter caps, naturally they should be changed, but no design change in the power supply area is necessary to get rid of the hum.  There are many DX-60's on the air with no hum and no changes in the P.S.!!  

Joe  

[KA2DZT]

You can measure the ripple with a scope or a digital voltmeter.  If you have a good DVM, measure your DC voltage, then switch the meter to AC and lower the scale down to a low scale (few volts). the meter should read only a very small % of the full DC voltage.  A supply with only one choke you may see 2-4 volts ac.  If you see a lot more, then you need to do more work on the supply.  A RX choke should be fine if it's a good choke.

To repeat what Stu said,  check the rectifiers to be sure both halves of the xfmr are working.  In fact, check the secondary winding to be sure one half is not open.  A half open xfmr will create lots of 60hz ripple.

Disconnect the feedback line from the 6DE7 to be sure the hum is not coming from that tube.

When you work on the hum problem disconnect everything from the mike connector.  Short the mike connector, any hum that still remains is originating in the xmtr.

I think you added a choke to the set, placement of a choke on the chassis can cause hum to be pick up in the audio.  Sometimes you need to rotate the choke 90 deg.  You can remove the choke and outboard it away from the chassis to test the affect the choke has on the hum, if any.

With everything that has been suggested, you should follow what Joe suggests and that a DX-60 can be made to operate on AM hum free.

The only thing you haven't told us is how are you listening to this hum,  hopefully not with a receiver placed very close to the xmtr.

Fred

[n1ps]

Lots of comments...

As I wrote in the initial post, I have both ripple (120hz is correct) and 60 hz components in the transmission.  I repeat...I replaced the entire power supply except the transformer.  This A model is a noisy beast.  Currently I have more than 5X the design capacitance with all new quality caps.  Yet the ripple is way too high (audible in a receiver). The gain control does kill the 60 hz but not the ripple.  I was more alarmed about the 60 hz because I cannot figure out where it came from.  I have decided to deal with the ripple first at this point.

As to changing the design, there is a dx60 mod on the AM window (posted by Steve I think) that shows replacing R34 with a choke, which is what I have in mind.  Worth a try isn't it?

TNX

p

[KA2DZT]

I think a voltage doubler is a 60hz power supply.  On half cycle the lower caps are charged, and the upper caps are charged on the other half cycle.

Never liked voltage doublers.

Fred

[n1ps]

Hi Fred again....I think it is a full wave...the lower caps get charged on the neg cycle and the top caps get charged on the pos cycle IIRC.  The scope seemed to indicate this... but dont take that to the bank. 

And I agree on the doubler comment.

Cheers
p

[AB2EZ]

I think Fred is right with respect to the 300V supply. I gets recharged at a 60Hz rate.

The top half of the 600V supply gets recharged in the first half of each 60Hz cycle. The bottom half of the 600V supply gets recharged in the 2nd half of each cycle...

But, the extra current drain from the 300V bus makes the amount of discharge and recharge bigger in the bottom half.

That makes the residual saw tooth waves on both the 300V bus and the 600V bus have a 60Hz repetition rate... with the 600V bus having an additional 120Hz sawtooth superimposed.

The bias supply employs a half wave rectifier... so any residual sawtooth variation of the bias voltage (probably very small) would have a 60Hz repetition rate.

It is the 300V bus (60Hz sawtooth)  that provides plate voltage to the low level audio stages.

If there is a strong 120Hz sawtooth component (superimposed on a 60Hz sawtooth component) on the envelope of the modulated output waveform, it could originate from plate modulation of the 6146 by the residual 120Hz sawtooth on the 600V supply or plate modulation of the 6DE7 by the residual 120Hz sawtooth on the 600V supply.

I wonder if there is 120Hz modulation on the envelope of the output power when the DX-60 is in CW mode (audio output stage disconnected from the screen of the 6146)? 

 

[KA2DZT]

Thanks Stu,

I was working on a project and thinking about this voltage doubler.  Came to the conclusion that the lower supply is charged at 60hz.  This half cycle pulse charges up the lower caps and also pushes up the voltage on the positive end of the upper caps to this same level.  On the next half cycle the upper caps are charge up by an additional voltage which adds to the lower voltage.  So, it seems the upper supply is also charged at a 60hz rate, but it may look like 120hz.  On the first half cycle the voltage on the upper supply is only the pulse voltage from the lower supply being transferred to the pos end of the upper cap.  This first half cycle pulse does not charge the upper caps it just biases them up to the lower voltage level.  On the next half cycle the upper caps receive their first charge pulse.  Both supplies charge at 60hz but 180degs out of step.

This may explain why the OP says he sees both 60hz and 120hz ripple.  Any ripple that appears on the lower supply will be transferred to the upper supply regardless of how big the upper caps are.  Any ripple from the lower supply will ride next to any ripple from the upper supply making it look like 120hz.  The greater the load is on the lower supply the greater the ripple.  It seems that a choke should be put in the lower supply to help reduce ripple.

The OP tried a choke in the upper supply, but if there is ripple on the lower supply, the ripple will be present on the upper supply regardless of the choke or how big the caps are.

Fred

It's now 3:30AM and I decided to take a look at Bill Orr's Handbook.  There are two types of voltage doublers, one is considered half wave and the other full wave.  As near as I can tell the one used in the DX-60 is full wave and the ripple freq would be 120hz.  I still think that each cap is charging at a 60hz rate, like I described above, but because they are 180degs apart the results look like 120hz.  This differs from a FWct or FWB supply were the caps are charging at a 120hz rate.

I think some of the confusion comes from the fact that most voltage doublers were used for only one voltage level and current is not pulled from the midpoint of the two caps.  Pulling current from the midpoint causes ripple to appear there at 60hz.

These supplies were used in TVs years ago with a 1-1 line xfmr.  In fact I still have a few of those xfmrs.

Now I know why I don't like voltage doublers.

Another point,  looking at the DX-60 supply, I'm not so sure that you can put a choke in one of supply line without putting another same choke in the other supply line.

Having said all of this, I still think that any ripple that's on the lower voltage line will appear on the higher voltage line regardless of the amount of filtering.  However, using a single cap from the high voltage line to ground would reduce this ripple.

[kb3rdt]

I have a DX-60!  The caps all beefed up has wc3k audio mods and ptt.

[AB2EZ]

Attached is the simulation of the stock DX-60 power supply. I set the input voltage to 120VAC (v. 110VAC in the manual/schematic). I added 100 ohms of effective transformer winding resistance. I set the 300V load to be 20mA (15000 ohms in addition to the bleeder resistor). I set the 600V load to be 125mA (4800 ohms).

Note that the 300V supply has primarily 60Hz ripple... but it also has 120Hz ripple.
Note that the 600V supply has a combination of 60Hz ripple and 120Hz ripple... in roughly equal parts.

The peak-to-peak value of the 600V ripple is 20V (3.33%), which would result in 1.67% peak plate modulation of the output of the 6146. This would be perceptible, but not very noticeable.

I wonder what the end-to-end DC resistances of the transformer secondary winding and (separately) primary winding actually are.

Stu

[KA2DZT]

Stu,

Those waveforms are just about what I thought they would look like.  A combination of both 60hz and 120hz.

Fred

I think, on the 300v waveform, the small 120hz bump comes from the fact that the lower caps get some small amount of charge from the charge pulse that occurs on the upper caps 180degs later.

Wondering what the waveforms would look like if you removed all load from the low voltage line.

[N2DTS]

Heathkit strikes again!

I guess that power supply design was typical for ssb stuff back then, but all those other supplies seem to work ok without a huge amount of garbage on the DC....

I have a drake supply, same stuff, and I expect the hp23 or whatever Heath supply that was used for the hw101 type stuff was the same, no?
I never had (a lot) of hum with that stuff.

[DMOD]

I have just upgraded another DX-60 (not DX-60B) and the power supply schematics are shown below:

At the junction of the bleeder resistor R20A and R20B (half HV) the ripple voltage is 1.5V P-P when keyed, 3.885 MHz, 20 Watts out in AM mode.

At the right side of the new 330 ohm 10W resistor the ripple is 1.0V P-P when keyed, 3.885 MHz, 20 Watts out in AM mode.

At the junction of R21/C33 (DX60B), R33/C32 (DX60),the ripple is 20mV P-P when keyed, 3.885 MHz, 20 Watts out in AM mode. The new value of C33/C32 is 47 uF.

The ripple on the 150V bias supply is 600mV P-P when keyed, 3.885 MHz, 20 Watts out in AM mode.

Now on the DX-60B, the bleeder resistors are on the diode side and the series resistors R34 and R35 are of unequal value. I would not replace R34 with a choke unless I also replaced R35 with a choke of the same value.

In fact, I prefer the DX60 power supply configuration because of the better regulation.

If you need some more voltages measured for comparison let me know.

Phil - AC0OB

[KA2DZT]

I have just upgraded another DX-60 (not DX-60B) and the power supply schematics are shown below:

At the junction of the bleeder resistor R20A and R20B (half HV) the ripple voltage is 1.5V P-P when keyed, 3.885 MHz, 20 Watts out in AM mode.

At the right side of the new 330 ohm 10W resistor the ripple is 1.0V P-P when keyed, 3.885 MHz, 20 Watts out in AM mode.

At the junction of R21/C33, the ripple is 200V P-P when keyed, 3.885 MHz, 20 Watts out in AM mode. The new value of C33 is 47 uF.

The ripple on the 150V bias supply is 250mV P-P when keyed, 3.885 MHz, 20 Watts out in AM mode.

Now on the DX-60B, the bleeder resistors are on the diode side and the series resistors R34 and R35 are of unequal value. I would not replace R34 with a choke unless I also replaced R35 with a choke of the same value.

In fact, I prefer the DX60 power supply configuration because of the better regulation.

If you need some more voltages measured for comparison let me know.

Phil - AC0OB

Thanks Phil for the measurements.  They look to be just about right for a power supply ripple.  I agree about replacing R-34 with a choke without doing the same on the low voltage line.  I wasn't sure but you say that R-34 and R-35 are not the same value,  this is problematic, they should be the same value.

Fred

I took a close look at the two supplies,  there is a major difference in the configuration of the series resistors and filter caps.  I'm thinking that the DX-60 layout is better because there is no current being loaded at the midpoint of the first set of caps.