Brain power needed

[kc2hgc]

Hi every one,

Steve here nice to meet some of you on the air by the way.
This may be of some  interest I am working on a collins 20v3 the bias supply is neg
120 volts dc   

No problem good 120volts    the problem is there is a 5volt pp ac riding on the dc 
This is not ripple it's an ac voltage I check every thing I even went as far as to change the
Transformer  and it got worse!! I isolated the supply from the transmitter and just ran the bias supply

This is a stander supply that has a 5U4 cathode grounded center tap  transformer  choke input
What do you guys think ?

Tnx.. Steve

[ka4koe]

Not familiar at all with this unit, but I would check all your filter components, i.e. chokes, filter caps, etc. How does this compare with what is specified, i.e. % regulation?

[kc2hgc]

Hi Steve here,

Yes the choke ,cap, wiring, check out I checked the transformer with ohm meter on 10meg
Looks good not leaking. 

This 5 volt ac is getting into the 400 grids 
Steve..

[KA2DZT]

How is the choke connected in the supply??

Is the supply a stock circuit or has it been modified in any way??

A 5U4 seems a little overkill for a 120V bias supply.  Are you sure the circuit didn't start out with a separate cathode type rectifier tube??

The bias voltage xfmr should have a CT HV winding where the neg 120 is delivered, the choke should be connected at this point.

Fred

[kc2hgc]

Hi Fred ,
We meet on the air I think last mount,

Ok yes stock circuit, center tap transformer choke connected to center tap 120 neg
 5U4 cathode grounded .
I checked transformer for leaking 10meg ohm digital meter look gook.
Maybe I can try a separate cathode tube.

This is strange  Where the hell is this 5 volts coming from?

[KA2DZT]

With the CT filament xfmr on the 5U4, half the filament voltage will always be riding on the HV leads from the xfmr.  I think the circuit used a different rectifier, one with a separate cathode.  Check this out and let us know.

Fred

[DMOD]

Hi Steve,

Whats. the frequency of this signal, 60 Hz or 120Hz? And are you measuring this to the right of C180?

Have you replaced C180? Do we know if it is any good?

The 20V3 has a separate filament winding on T105 for the 5U4 rectifier tube.

Phil - AC0OB

[kc2hgc]

Hi Phil nice to meet you

It looks like 60hz I will put the freak meter on it tomorrow ,
measurements are at pin 6 off of tb 108 I found this when I was checking the 807 driver
Output .

It's making a mess at the 400 grids..

Tnx Steve.

[W7TFO]

Check the 4-400 filament circuit, sounds like a bit of hum is being introduced there.

That circuit should be balanced with a center tapped fil trans.

73DG

[AB2EZ]

Steve

From the schematic of the 20V3:

One side of the 5U4 filament is connected to chassis ground at pin 8 of the 5U4 socket. The positive side of the filter capacitor is also connected to ground via pin 8 of the 5U4 socket.

Make sure that both of these connections to ground are good... by measuring the voltage between the positive side of the filter capacitor and ground (should be zero volts AC and zero volts DC).

If the positive side of the capacitor were mistakenly connected to the non-grounded side of the 5U4 filament, then the AC filament voltage would add to the intended DC bias voltage at the output of the bias supply.

Stu

[KA2DZT]

Stu,

I think the circuit is using a CT filament xfmr, with the CT grounded.  I think this can be a problem.  All the 5U4s I ever saw, the DC was pulled from one side of the filament, pin 2 or 8.  Pins 4,6 are the plates.

Steve claims the circuit is stock, but I'm not so sure of that. I think something has been changed.  I'm not familiar with the 20V3, but the use of a 5U4 for a 120v bias supply seems a bit overkill.  I suggested a separate cathode type rectifier.

In a few PMs from Steve earlier today, the first thing I told him was that there is something wrong with the grounds or the method he's using to measure the AC voltage.

Your point of the filter cap being connected to the wrong ungrounded pin of the filament is also possible.  Or, the cap is connected to one filament pin while the CT of the filament xfmr is being used to make ground.

Fred

[AB2EZ]

Fred

et al.

The 20V3 schematic is available, on line. What it shows is: a 5U4 rectifier for the bias supply, with the filament winding of the bias supply transformer being connected to ground on one side (no filament winding CT).

In this application (a full wave rectifier), it doesn't matter if the filament is balanced with respect to ground. [Unlike the filament supply for an amplifier tube whose cathode is also the filament or directly connected to the filament].

What does matter (in this application) is that the positive side of the filter capacitor is making good contact directly to ground.

As a separate issue: it is important that the filter capacitor is good (i.e. has its full capacitance); and it is also somewhat important that both halves of the 5U4 are good. If one side of the 5U4 is not good, or if one plate of the 5U4 is not properly connected to the transformer.... then the residual hum at the output of the bias supply will be around twice as large as normal; and also the residual hum will be at 60Hz rather than 120Hz.

Stu

[KA2DZT]

Stu,

Thanks,  funny, none of these problems are mysteries,  If you look long enough and far enough, you'll find the answer.

We'll wait to see what Steve discovers, hopefully something.

Fred

[Pete, WA2CWA]

For those that are head scratching, here is the schematic in pieces:
http://collinsradio.org/archives/manuals/20V-3__09-61_-Sec7.pdf

[WU2D]

60 Hz also could be the AC for the finals themselves which are also directly heated which would ride right on the bias and get coupled right up to the grids. They knew about hum since they put nice hum nulling pots on the modulators and finals fils.

[w1vtp]

Hi Phil nice to meet you

It looks like 60hz I will put the freak meter on it tomorrow ,
measurements are at pin 6 off of tb 108 I found this when I was checking the 807 driver
Output .

It's making a mess at the 400 grids..

Tnx Steve.

On the scope, 60 Hz = 16.7 ms, 120 Hz = 8.3 ms per complete cycle.

[WU2D]

Al that Hz.

[kc2hgc]

Hi guys,

up date well I cant find Y or were the 5 volts is coming from maybe phase relation to cathode,
transformer?

However By trashing the 8 mf oil cap and adding two 50mf electrolyte one on each side of the choke
cleaned up the ac down to 50mv and none traceable on the dc super CLEAN
 
I will have to adjust voltage divider as there was a raise of 30 volts.
let me here what ya think not sure if its the right fix but it WORKS!

Thank`s guys for all the help! now on to the RF & AF Drivers 73 Steve       
 

[DMOD]

Hi guys,

up date well I cant find Y or were the 5 volts is coming from maybe phase relation to cathode,
transformer?

However By trashing the 8 mf oil cap and adding two 50mf electrolyte one on each side of the choke
cleaned up the ac down to 50mv and none traceable on the dc super CLEAN
 
I will have to adjust voltage divider as there was a raise of 30 volts.
let me here what ya think not sure if its the right fix but it WORKS!

Thank`s guys for all the help! now on to the RF & AF Drivers 73 Steve        
 

I don' know what you mean by phase relation to cathode transformer.

As Stu mentioned, according to the schematic the filament transformer is a single-ended 5V secondary on T105 with one lead connected to ground, and the other connected to one side of the 5U4 filament. The other side of the filament is grounded.

You never did tell us where you were measuring this 5 VAC signal.

Were you looking at this 5 VAC on an oscilloscope with your probe connected to the right of L112, the filtered side?

Earlier suggestions were to check the value of the 8 uF oil cap to see if was actually 8 uF. Personally, I have never seen an 8 uF oil cap at that voltage go bad.

Phil - AC0OB

[kc2hgc]

Hi Phil ,
I was measuring at the supply output at tb 108 with the supply totally disconnected ( pin 6)
Ac was ridding on the dc.

As for the phase thing I ment to say possibility some type of phase relations to the cathode via the transformer  I really can't finger this out.or maybe a wiring mixup?

[kc2hgc]

Oh yes I was looking at this on a scope 5volts p-p ac on the bias supply ac was ridding on the dc and when looked at with scope set on dc you see 120volts dc with ac

Again I can't finger it out

[KA2DZT]

You mentioned in one of your earlier post about replacing a xfmr, what xfmr??  OK on replacing the cap, not so sure about adding the second cap on the input side of the choke.

As mentioned, the filament xfmr for the bias 5U4 should NOT be a CT type xfmr.  One lead of the filament xfmr is grounded, one filament socket pin is also grounded.  The positive cap lead is also grounded at the grounded socket pin.

You need to check all these grounded connections to be sure the grounds are making solid connections, especially the grounded cap connection.

Another thing to check is the 5U4,  I've seen a number of 5U4s with cold solder on the tube pin.  Remove the tube and look at the end of the filament pins to see if they're still well solder.  This solder issue would not cause you problem but make the 5U4 to not light.  I thought I would just mention it.

Fred

[AB2EZ]

Steve

Here is what I think happened:

The positive side of the capacitor was connected to the hot side of the 5U4 filament instead of the grounded side. This could have  been caused by one of several possible wiring errors. This would cause the AC filament voltage to add to the negative DC voltage at the output of the bias supply.

When you removed (or disconnected) the stock 8uF capacitor, you removed this wiring error.

When you added the new capacitor at the output side of the supply, you correctly connected the positive end of the new capacitor to the nearest convenient ground.

This fixed the problem.

One capacitor, on the output side of the choke is sufficient and preferable. [I.e. as you observed, adding a capacitor to the input side of the choke creates several problems, including increasing the output voltage of the bias supply.]

If you took out both of the new capacitors, and properly reinstalled the stock 8uF capacitor (with the positive side correctly connected to the nearest ground), the added AC on the output of the bias supply would be gone.

Stu

[KA2DZT]

Stu,

I agree with everything you say.  Even ground the cap right at the 5U4 tube socket (grounded pin).

Fred

[AB2EZ]

Here is an interesting scenario, but not the most likely cause of the problem:

1. The connection between pin 8 of the 5U4 socket and ground becomes defective (high resistance)

2. As a result, both the filament and the positive side of the capacitor lose their connection to ground

3. The filament goes dark

4. Someone tries replacing the 5U4, but that doesn't fix the problem

5. He assumes (incorrectly) that the transformer's filament winding is open

6. He adds a pair of 1N4007 diodes to the socket of the tube, and he leaves the tube in place (still dark)

7. The bias supply works again, except the path between the positive side of the capacitor and ground is through the 5U4 filament (not lit, but the tube is plugged in) and through the 5VAC winding of the transformer

8. The result is 5VAC riding on the DC bias

Stu