Author Topic: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement  (Read 5429 times)

0 Members and 1 Guest are viewing this topic.

Offline factory

  • Super Contributor
  • ***
  • Posts: 3215
  • Country: gb
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #50 on: November 14, 2024, 08:05:23 pm »
If the problem is vertical deflection I would be looking at the vertical circuitry. HV fault would affect the horizontal deflection, trace usually gets wider/longer if HV drops.

As for measuring the HV, are there no AVO lamp foots over there? some of the model 8 have a 3kV range, probably enough for this as the HV isn't that high compared with higher frequency scopes. Only problem in finding good one, took me a few attempts, they go for about £10 here.

David
« Last Edit: November 14, 2024, 08:13:26 pm by factory »
 

Offline dazz1Topic starter

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: nz
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #51 on: November 14, 2024, 09:18:18 pm »
Hi
Yes, Avo's are available here but they are now rare enough to sell for prices 10x to 20x your £10. 


My highest rated meter is a Fluke 77 @ 1kV so well short of being useful.  Even if it was rated higher, the 10Mohm meter input impedance would load down the 3kV supply.


The HP 3721A has a 3kV supply with a tolerance of +/- 2v.  That tolerance will provide accurate measurements off the CRT image, but it actually makes more sense to read the data from the pre-GPIB data bus that uses a 50pin Centronic connector.  Rationally I should ignore the CRT fault and spend my time and money emulating a HP 3720A Spectrum Display.

I think I have been dealing with at least 3 separate faults.  There was the carbon compound resistor identified by TauTech.  Cleaning the corrosion of the CRT pins seems to have fixed the Y deflection modulation by the brightness control, and apparent breakdown of a mystery HV component.  None of the low voltage things I can measure correlate with what I see on the display.
Dazz

Over Engineering: Why make something simple when you can make it really complicated AND get it to work?
 

Offline dazz1Topic starter

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: nz
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #52 on: November 15, 2024, 12:26:06 am »
Hi
So while a rational person might just leave the CRT fault as is, I never claimed to be rational.

I took a harder look at the circuit diagram to figure out where a fault could exist that did not generate an error signal.  I narrowed it down to R17, a 1Mohm LV resistor.  When measured unpowered, this resistor had and does check out OK.
So I changed that resistor out and changed the fault.
The time constant significantly reduced and the Y-deflection went full scale.  The unstable trace is now stable.   In addition, adjusting the intensity no longer has a significant effect on the Y-deflection.   Almost BINGO.
The photos show the before and after changing R17. 

Based on these observations, I now believe that R17 was failing to a higher resistance state that created a significant time constant re-charging C7 and maintaining the grid voltage. 
I now suspect that C7 has or is failing to an open circuit.  Without the full capacitance of C7, it takes time for the current to flow through R17 to build up the voltage on the control grid of the CRT.

My next problem is obtaining a new HV cap.  I just can't wander down to the local corner shop and pick one off the shelf.  I have to wait until I have a big enough parts order to take advantage of the free shipping.   Typically I order parts about twice a year.  I end up with multiple projects on hold until I get parts to progress them.  So expect radio silence on this topic for quite a while.
« Last Edit: November 15, 2024, 12:28:08 am by dazz1 »
Dazz

Over Engineering: Why make something simple when you can make it really complicated AND get it to work?
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 29816
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 

Offline dazz1Topic starter

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: nz
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #54 on: November 15, 2024, 04:38:42 am »
My next problem is obtaining a new HV cap. 
2m search:
https://www.aliexpress.com/item/1005006067184429.htm

I buy a lot of stuff of Aliexpress but I hadn't considered buying HV caps.  I will look into that.
Dazz

Over Engineering: Why make something simple when you can make it really complicated AND get it to work?
 

Offline dazz1Topic starter

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: nz
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #55 on: November 16, 2024, 04:41:07 am »
My next problem is obtaining a new HV cap. 
2m search:
https://www.aliexpress.com/item/1005006067184429.htm

I buy a lot of stuff of Aliexpress but I hadn't considered buying HV caps.  I will look into that.

I have looked into the Aliexpress option for HV caps and decided not to go that path.
I can buy 5x HV caps from Aliexpress for the price of 1x HV cap from element14. 
I only need 1x so buying 5x caps is not an advantage. 
I would have greater confidence in the quality of parts from element14, so the plan is to buy from element14. 
That won't stop me from buying a lot more stuff from Aliexpress. 
Dazz

Over Engineering: Why make something simple when you can make it really complicated AND get it to work?
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 29816
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #56 on: November 16, 2024, 05:07:04 am »
My next problem is obtaining a new HV cap. 
2m search:
https://www.aliexpress.com/item/1005006067184429.htm

I buy a lot of stuff of Aliexpress but I hadn't considered buying HV caps.  I will look into that.

I have looked into the Aliexpress option for HV caps and decided not to go that path.
I can buy 5x HV caps from Aliexpress for the price of 1x HV cap from element14. 
I only need 1x so buying 5x caps is not an advantage. 
I would have greater confidence in the quality of parts from element14, so the plan is to buy from element14. 
That won't stop me from buying a lot more stuff from Aliexpress.
Yet there are 5x 4700pF caps in that HV circuit......

Back when I repaired CRO's I sourced all the HV caps required from Ali....
YMMV
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 

Offline factory

  • Super Contributor
  • ***
  • Posts: 3215
  • Country: gb
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #57 on: November 16, 2024, 12:07:20 pm »
Trouble with parts from aliexpress is you never know if they will meet the spec (if there is a spec at all, can't view ali for some reason), they may send good parts one day, crap the next time.

If you haven't got a way of measuring 3kV, then I'm guessing you probably don't have a way of testing HV disc capacitors either. One thing to note with ceramic disc caps, is the capacitance changing with voltage applied, the capacitance drops with increasing voltage.

David
« Last Edit: November 16, 2024, 12:09:46 pm by factory »
 

Offline dazz1Topic starter

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: nz
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #58 on: November 17, 2024, 09:54:36 am »
Hi
If there was a lot of $ at stake here, I would definitely give Aliexpress a go. The key to successful buying off Aliexpress is to buy the most expensive version, which is invariably heaps cheaper than a locally available version.  With Aliexpress, more money = more quality.   In this case, there is no cost saving compared to element14 because either way, I still have to pay out about the same amount of $.

There are multiple HV caps of the same value/voltage on my HP 3721A, but only one has failed.  The failed one has led an arduous life because it has pulse current due to the 100 blips on each x deflection.  The other HV caps basically sit at DC.

Another issue is the physical size.  The existing caps have a pin spacing of about 15mm, but replacement options have spacing down to 5mm.  I don't think 3kV across a 5mm gap is a recipe for long term reliability because of the risk of dirt leading to leakage, or even arc-over.

With no cash incentive to buy Aliexpress, the assumed quality and known physical dimensions of the element14 range win the day. 
« Last Edit: November 18, 2024, 01:34:01 am by dazz1 »
Dazz

Over Engineering: Why make something simple when you can make it really complicated AND get it to work?
 

Offline dazz1Topic starter

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: nz
Re: Vintage HP 3721A Correlator : New knobs
« Reply #59 on: December 06, 2024, 09:14:51 am »
Hi
I figured it would be too difficult and too expensive to replace the missing and broken original late model HP knobs.     I would have to buy from multiple suppliers and the total shipping costs would eye-watering.  Also genuine HP knobs only seem to be over-priced.  The broken and missing knobs were really annoying me.  After a lot of research I found some close equivalents to the old model HP knobs.  These work better than the late model knobs because the front panel art work was made for them.  The late model knobs are a little too small.

So the new knobs were not cheap and cost me twice as much as I paid for the HP 3721A but I think they are worth it.

I had to modify knobs to fit the double decker control switch "Delay Offset" but the finished job almost looks factory.  I am going to paint the pointer line and the brushed aluminium bits blue to match the panel.

I have attached a photo of an original HP with black knobs and my copy of the HP black knobs. 
« Last Edit: December 06, 2024, 09:22:37 am by dazz1 »
Dazz

Over Engineering: Why make something simple when you can make it really complicated AND get it to work?
 

Offline ch_scr

  • Frequent Contributor
  • **
  • Posts: 885
  • Country: de
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #60 on: December 06, 2024, 10:16:25 am »
Great job, looks factory  :-+
 
The following users thanked this post: dazz1

Offline dazz1Topic starter

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: nz
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #61 on: December 07, 2024, 07:40:44 pm »
Hi
Now working on improving the thermal management.  Cooling might have been sufficient in the chill of Queensferry in the north of Scotland, but not so good in a temperate climate.  It is summer here.

I have fitted heat sinks to the TO-3 power transistors on the rear.  The aim being to dump heat outside the enclosure before it gets into the chassis.  The heat sinks will fit under the factory cover.
I have also fitted a thermostatic controller to the fan so it only runs as fast as necessary.

I have also replaced the HV cap but now have an issue with power supply fuses blowing.   I can't yet test the fix.   Before you ask, the heat sinks are electrically isolated as they should be.
The problem is on the 5V supply which draws a LOT of power on this thing.   One of the stud diodes on the 5V rectifier has previously been replaced so I am not the first to have a problem.

There is a before and after fitting the heat sinks.
Also a photo showing the fan controller fitted to a 3D printed bracket, wiring not yet fully loomed-in.
« Last Edit: December 08, 2024, 10:37:36 am by dazz1 »
Dazz

Over Engineering: Why make something simple when you can make it really complicated AND get it to work?
 

Offline dazz1Topic starter

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: nz
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #62 on: December 11, 2024, 07:08:24 am »
Hi
I went through and did a lot of static checks on the power supplies to make sure there were no obvious broken parts. 
The power supplies have a cascading protection scheme starting with the 50V supply.  The +24VDC supply requires a 50V input to the reference.  If the 50V supply fails and goes to zero, the 24VDC reference is zero and the 24VDC supply goes to 0VDC. 

The +12VDC supply reference also depends on the 50VDC supply.  The +5VDC supply reference is driven by the +12VDC supply.  Other supplies also have dependencies.

As a result of all this interconnection, if one fuse blows, multiple power supplies drop out.  It looked really bad, until I dragged out the schematics to see how everything works.  Once I had completed static checks of the key components in the power supplies, I refitted/replaced the fuses as required, and flicked the power switch.  It all burst back into life again as if it was just another day. 

The +5VDC power supply has some interesting features, if you are a power supply geek.   It runs 2x 2N3055 power transistors in parallel because one was obviously not enough.  That is a lot of power.   The transistor that supplies the base currents is itself a power transistor.  Being all linear power supplies, there is a lot of waste heat.

The newly fitted heat sinks are definitely doing their job.  The heat sinks combined with the fan controller are working well together.  I can't quantify how much heat is being dumped by the heat sinks, but it is significant.    I have tuned the fan controller to start accelerating from a low base speed at 25 deg C, and go to full speed at 50 deg C. 

With the power supplies now fully operational, I checked the repair to the CRT power supply (new HV cap).  The good news is that the symptoms consistent with a failed HV capacitor have gone.  The bad news is that the power supply output is unstable, consistent with arc-over somewhere.  That brings the tally to 4 defective/replaced components and at least one more to discover.  I now have a used HV 1000:1 probe so I can now poke around to see new things I could not see before.

Three of the rotary switches were loosely mounted, so I tightened those up.


To the HP employee that left the message "FIXIT" 55 years ago please note, where ever you are, I am trying. 

So the next step is to figure out where the remaining HV fault(s) is.  So far no magic smoke has been observed during test flights.



« Last Edit: December 11, 2024, 07:46:24 am by dazz1 »
Dazz

Over Engineering: Why make something simple when you can make it really complicated AND get it to work?
 

Offline dazz1Topic starter

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: nz
Re: Vintage HP 3721A Correlator : Repair, Restoration and Enhancement
« Reply #63 on: December 13, 2024, 05:09:39 am »
Hi
Photo below IMG_3185 shows the crt display before the HV cap and resistor were replaced.  The cap value had degraded.  The "dot" are near vertical as the HV supply dropped with the increased current required to draw a dot.


Photo IMG_3316 is the after the fix.  The dots are closer to being dots.  Vertical deflection is still about half what it should be so still a fault there to fix.  There is no droop in the line of dots.  The line is straight.
2460399-0
« Last Edit: December 13, 2024, 05:16:56 am by dazz1 »
Dazz

Over Engineering: Why make something simple when you can make it really complicated AND get it to work?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf