HP 3562A Signal Analyser with broken power supply : need help

[psykok]

Hi,


I just got my new Signal Analyser, a nice HP 3562A ... a big beast.
I bought it as broken, the seller described as : instrument with shorted psu.


The first issue was to find a complet service manual with the psu section.
All 3562a pdf we can download and specialy all from HP/Agilent are not complet.
By chance the 3563A shares some component with the 3562A adn I found all necessary information about the psu in the service manual of the 3562A.

The service manual PSU section:
http://alexkachler.free.fr/workshop/3562a/doc/3562a_psu.pdf


I started my investigation the last week and found out that the PRI (CR6 led)debug led was on.
Based on the service manual the Primary is faulty

I  started with the Primary Fault Test but I had o stop on point 9 because I have currently no VARIAC.
But until this point every thin was correct:
 - Q400 and Q401 :  ok.
 - R409, R410, R403 and R404 : ok.
 - Waveform #4 : ok

I have checked some other waveform I found inthe manual:
 - U2 : Wavforme #1 :ok


 - Driver and PWM Clock : Wavforme #2 :ok


 - Sync : Wavforme #3 :ok


 - Test Switching Waveform  : Wavforme #4 :ok



 - power FET Gates : Waveform #1 : strange results



 - Transformer T2 : Waveform #6 : FAILED



Is the T2 transformer broken or a side component?
Do you have an idea how I can continue? or what can be broken?

Thanks for your help


Regards,

Alex

[Richard Head]

Alex

It looks like you are running the PSU with no load as there is no output from the CT.
I sugest you load the main output to perhaps 50% (if possible) and look at the CT (T2) output waveform. It should be a squarewave with a trapezoidal ramp on both sides. (Difficult to describe!)

[psykok]

Hi,

No the PSU is loaded, there is a jumper on the board (A18J100) to have a load during the test.

To be sure I will check it again if I had this jumper in place during my tests.

Alex

[Richard Head]

Incidently, your gate drive waveform looks weird because your ground lead is on the wrong side of the gate drive xformer. Also, don't try to connect the ground lead to the high side MOSFET source as you'll blow up the PSU. You can connect the ground lead to the source of the lower MOSFET without a problem.
If you have an isolated scope system like the Tektronix P5200 or similar you can view both gate drives at the same time but I'm sure the gate drives are fine in any case.

If the PSU is loaded then is may be that your CT has a short circuited turn on the secondary.
What is the actual problem with the PSU?

[psykok]

Hi,

I'm always very careful wen I have to measure something with my scope in a PSU and specially with switching PSU.
In this case I have only appalled the measurement procedure described in the service guide.


The solution was to connect CH1 to TP8 and the ground on R800 (the red point)


Unfortunately I have no isolated scope probe 

The problem is that the psu has not power after the main transformer T1.
Primary fault led is on.
Only the bias power supply which is directly connected to the main is providing the +5B



Alex

[oldway]

Check the primary of T1 for continuity and replace both 2µF capacitors C402 and C406.

[psykok]

C402 and c406 are the two first capa I have checked.
Both are ok,  right value and good esr arround 0.1R

Il will check if i found some continuity on the primary of T1

Alex

[psykok]

Hi,

Just to be sure I checked again c402 et 406 both are ok.
I made also a test with 2 new capa and same result as with the original one.

Alex

[oldway]

Dc voltage on each of those capacitors must be 1/2 of the full DC voltage....Have you checked this ?

Could you also measure the resistance between the center point of those two capacitors and the commun point of R400 and R408 ?
How much ohms ?

[psykok]

The res is arround 0.5R

I have 326v as full dc and 170v on c406 and 158v on c402

Alex

[oldway]

Replace the 2 Mosfet's Q400 and Q401.

[psykok]

Replace the 2 Mosfet's Q400 and Q401.

You think that he mosfets are dead?

I checked it yesterday with my DMM and everything lokked good.


I have to find some to3 mosfet equivalent to this hp part .... Not sure I have someting in my stock.

Alex

[Richard Head]

Measure for voltage before the post regulators on the output. Then, with a scope, check for a square wave output from the transformer secondary. If there is still nothing then measure for secondary continuity, then the primary.

[psykok]

Hi,

What do you mean with post regulators on the output? the series of LM337 or whatever we have at the output of T2?

For the measurement I will do it this afternoon.
The only think a can say is that the following test failed also:

I have no picture from the results but I will take it.


Alex

[psykok]

Hi,


I have made a test with 2 new mosfet and same results as before ...  the mosfet are for me ok.

I measured the voltage at the input of the regulator behind T2 and I got 3v wen I switch the power on and after that the voltage is decreasing until 0v


Alex

[Richard Head]

Psykok
The half brige converter isn't a conventional configuration. The converter runs wide open (50% duty cycle) with no voltage feedback. Its purpose is to provide galvanic isolation and supply multiple outputs to each linear regulator to reduce dissipation. There isn't a conventional voltage feedback loop as one might expect. The reason is that with a switchmode converter with multiple DC outputs it's very difficult to ensure good voltage regulation for each output. This is refered to as cross regulation. If you require good regulation on all outputs regardless of load then you require post linear regulators, as they have done.
Back to the repair though. As I suggested, measure the input voltages to the post regulators to confirm that the power is at least making it across the power transformer. If there's no input voltage to the regulators then measure the secondary of the HF transformer to see if anything is coming out of it.

[psykok]

Hi Richard,

Thanks for the explanation.

Like I said in my previews post, I measure the voltage one some post regulators and I got something around 3V but very shortly after powering the psu on.
The voltage come quickly up to 3v and decrease directly until 0v, what make sans ... the power down circuit is operating.

I tried to measure something on the output of HF transformer ... but I can't get anything, I think the time were we have something at the output is to short.

I got my variac yesterday, so I will be able to perform the rest of the test provided by HP for the primary section.


Debug switching psu is always a pain in the ass  for me, time consuming and and hard to resolve


Thanks for your help

Alex

[Richard Head]

Psykok

It seems something is open circuit. Maybe the xformer but that's not too likely.

Good luck. I'm sure you will find the fault quite soon.

[oldway]

Waveform #4 (Test Switching Waveform) is ok, that's mean that (as you said!): "I started my investigation the last week and found out that the PRI (CR6 led)debug led was on.
Based on the service manual the Primary is faulty"
That's right : 

In the primary, we have allready checked:
- C406 and C402 : ok
- continuity between center point C406/C402  and center point R408/R400 : ok
- Mosfets Q401 and Q400 :ok

We still have to check :
- C403 and C404
- Drive signal between Source and Gate of Q400 and Q401. For this, you have to use an isolation transformer and use ONLY one probe.

[psykok]

Hi,

The primary faulty message can also be activated by a faulty Power Down section ... I will check that today afternoon with the variac.

Coming back on the primary, I have check the  big capa C403, C404 and both are ok if I trust my LRC meter
I had some doubt about C403 because the capa box has a smal hit, but value and ESR are in specs.

How can you proceed to measure the SG signal on Q400 and Q401? which kind of transformer should I use?
Not sure I have the necessary hardware to do that.


Alex

[oldway]

 Waveform #6 : FAILED (Transformer T2) show that there is no current circulating in the primary although waveform 4 is OK.
Why is this happening ?
- or something is interrupted in the path of the current,....whe have already checked, everything is OK.
- or, Q400 and/or Q401 are not switching.
Q400 and Q401 are OK (already check)...the only possibility I see is that one of them does not have the correct gate drive signal.
Perhaps one of the diodes open  (CR700....CR707) or one of the secondaries of T4 open ?

To measure de gate waveform on the main side of the power supply, you need a differential probe or to isolate the power supply from the main using a 1/1 isolation transformer.

[psykok]

Hi,

I'm back and with a good news .... I found the faulty component.
The big 80A diod CR300 was shorted
Unfortunatly I don't have a diode with equal value but by chance I found a 40A same package.
After changing the diode the faulty message disapear and I have again all right voltage value at the ouput.

Now I have to order the right spare diode and remonte everything to do further tests.

Again thanks for your support !

Alex

[psykok]

Hi,

The story is not finished.
Yesterday I had to exchange the rectifier bridge in the bias power supply.

The  pulse modulator SG3525 in the screen psu was also dead ...

Unfortunatly the psu is also not stable in. Charge after 1 min or less, it' turning off.
The 5v is only 4.7 what is in specs but a little be strange for me.
I thing I still have something faulty in the 5v section.


Alex

[Richard Head]

Psykok
I'm glad to hear that you have found the problem. I just wonder why you didn't see current overshoot on the CT output waveform since one of the output diodes is short circuited.

[psykok]

Hi,

I don't know why I don't see that.

Now i have to wait for the spare parts.

Alex

[psykok]

Hi,

I got the first replacement part: the big diode:



After replacing the diode, I turned the PSU on and again a failure but an other one.
This time the bias part was broken, I had to replace all +/-12v regulator.

Now everything is again back, I have all output on the right value and the 5v is 5.1v and no more 4.7v.

The strange thing is that the psu is making some little whistling sound ... normal or bad sign?

Alex

[Richard Head]

If it's whistling I would put a scope on the DC output to see if it's a voltage control loop instability or just noise in the control loop.
I presume you used exactly the same output diode?

[psykok]

If it's whistling I would put a scope on the DC output to see if it's a voltage control loop instability or just noise in the control loop.
I presume you used exactly the same output diode?

Yes it's exactly the same diode.

OK will check it with the scope, I had no time yesterday evening to do more investigation.
For me the noise is coming from T3 in the bias power supply section, I have to check it to be sure.



Alex

 

[perlmonkey]

Did you make any progress on this? I have a 3563A which I bought "as is" from eBay; resurrected the CRT but the 7805s keep blowing and it has a ROM issue.
I also have a 3562A which was working fine... then I moved to a new house and now it doesn't work; all LEDs light up but don't turn off, no CRT now. Sigh.

[psykok]

Hi,


I don't made any progress on this 3562A, the power supply is working (I have all right voltage at all outputs) ... only in test mode (without charge)
If I charge the psu the power diode (the one I replaced) is getting very hot until braking.

Unfortunately after many ours investigation I still not found any solution.
The only evidence I have is the strange whistling sound coming from the primary transformer.

I have also fixed the broken crt on this instrument, a transistor was broken on the psu board.

By chance my second 3562A is working pretty good without issue


Did you checked the psu on you 3562a to be sure that the startup signal generated by the psu is present?



Alex

[katzohki]

...
The only evidence I have is the strange whistling sound coming from the primary transformer.
...

Uh oh. Could be arcing internal to the transformer.

[psykok]

...
The only evidence I have is the strange whistling sound coming from the primary transformer.
...

Uh oh. Could be arcing internal to the transformer.

Yes it could be
Honestly the transformer is one of the last component which I have not realy tested in this psu.

Is there any way to check if this transformer has a problem ?

Alex

[katzohki]

...
The only evidence I have is the strange whistling sound coming from the primary transformer.
...

Uh oh. Could be arcing internal to the transformer.

Yes it could be
Honestly the transformer is one of the last component which I have not realy tested in this psu.

Is there any way to check if this transformer has a problem ?

Alex

Well the easiest would be to compare a known working one. See if the output voltages match, winding inductance and resistance etc... Or even better, swap it. If it's arcing through when running it might not look like a short when testing at low voltages. Basically you need a Megger or HiPot tester I think.

The other thing to do I guess would be to take it out of circuit and simulate its output. Or at least simulate the various DC power lines.

If you can find out all the parameters of it from the service manual I guess it could be replaceable.

[almstsobur]

     Hey guys, my first post on the EEvblog, been a browser and YouTube watcher for a while. But I thought I needed to chime in here. I ran across this thread searching for a solution to just about the exact same problem, my 3562a also fails the same 3 scope patterns. I also think I have at the least an issue with the same 2 diodes, at least that is where my troubleshooting has ended until those parts get in. However, if someone else besides myself runs into this thread, like I did, I'm pretty sure you have the wrong diode.

      It looks from the picture that the diode you selected is the 85HFR120, however the original diode is a Schotty, also the forward voltage is 600-800mV on the original with the HFR being 1.2V, most of all I have attached a picture of my original and on the picture you have posted, the anode/cathode are in the wrong direction. Please correct me if I am wrong, but I hope this helps the original poster, or maybe someone who comes along this thread later. I have ordered the MBR7560 by GeneSiC Semiconductor as I think it is almost an exact match.

-RP

[almstsobur]

Just a quick follow up; the MBR7560 Diode's did infact work perfectly and my 3562A is fully functional again.

-RP