Tektronix SC504 HV multiplier

[Doug Brewster]

I'm in the process of dis-assembling a bad HV multiplier out of my SC504 to try and discover what's in it. I'm about 3mm in and have encountered what appears to be a couple of caps and the ground connection. What I'm looking for is info from any previous efforts in this forum to repair these devices in the hope it will save me work; the 3mm progress I've made so far using engineer's burrs in a dremel type drill has cost me an hour of dusty work so far and of course the damage done can make the internals unrecognisable as I found when doing the same job (happily successfully using info from the forum) in my 475, actually a couple of them.

Any ideas where to look? I'm new to this forum and haven't found my way around yet...

[floobydust]

I don't see anyone taking these apart and reverse engineering them.
Tek465 HV multiplier replacement https://www.eevblog.com/forum/repair/tektronix-465/msg672083/#msg672083

All you need to know is how many stages, and capacitor value for the HV multiplier, in 152-0634-00 or CMX222.
I've used a heat gun to basically roast epoxy potting compound and soften it enough to pry off with a screwdriver, to read component values.

Most are 4-6 stage, 1nF or 1.4nF 10kV, BY8212 Philips diodes 12kV 5mA.

I would use 2CL74 (14kV 5mA) to 2CL77 (20kV 5mA) fast-recovery HV rectifier diodes and 2,200pF-10,000pF caps.

There's some x-ray pics, (URL ending in pic #1-6 of link) of a Tek 465 multiplier insides.
The HV connector is the valuable part to salvage.

edit: fixed url

[Doug Brewster]

Thanks floobydust for the response. I  think it's a triple multiplier based on the SM info of 3.3KV pk input and 10KV (undefined) anode voltage. Trouble is the housing is about 1 cubic inch in size and it's going to be pretty crowded in there to get the parts in. I haven't tried the heat gun approach, maybe I'll give it a shot...

[chris_leyson]

U1590 152-0634-00 VOLTAGE MULTR:HV MULTR, 6.7kV IN, 10kV OUT. Description in the service manual reads,
"The accelerating potential is supplied by voltage tripler U1590, having an input voltage of approximately 3.3 kV peak."

So the "6.7 kV IN" in the part description is pk-pk voltage. Same multiplier is also used in the SC502.

[David Hess]

Annoyingly depending on the documentation, Tektronix might refer to the same high voltage multiplier as 1.5x or 3.0x.

[Doug Brewster]

So the "6.7 kV IN" in the part description is pk-pk voltage. Same multiplier is also used in the SC502.

Thanks for the responses. I've measured the input side at 1400V (unloaded) on my rms reading fluke which gets me nearly 4KV peak. My version of the SM doesn't make reference to 6.7KV. This uncertainty about the input voltage is what's making me be careful in the disassembly so that I can determine the number of caps and diodes present....

[Doug Brewster]

Been trying to post pics of progress; can't make it work and can't find any help on the topic. Somebody point me the in the right direction?

[floobydust]

Did you mean posting pics is not working for you? I use" Attachments and other options".  Some people link to image hosting sites.

[Doug Brewster]

It's "Attachments and other options" that I can'get to work. The image number appears in the "Choose a file" box when I select a pic but nothing else happens..

[Gyro]

Are you paying attention to the size limits? 1MB per image. Max 2MB per post. Try resizing.

P.S. I've been using 'Resize' since Win98 days. Simple, quick and good quality output, single .exe ... https://resize.en.softonic.com/

[tautech]

Been trying to post pics of progress; can't make it work and can't find any help on the topic. Somebody point me the in the right direction?

From yesterday in another thread:

It's not too hard once you know how.
Any one file can't be larger that 1 Mb.
You're only allowed a total of 2 Mb/post.

Use your photo editor of choice to compress pics to a size relevant to the detail you need to show. 100 KB is usually heaps enough detail.
Add them individually into the 'Attachments" UI below this posting box but take note of the file types allowed.

Then they're displayed as thumbnails that you can copy their URLs and place between in [ img ]image[ /img ] tags available from the Insert Image icon.

Sometimes for a few images it's easier to open dual tabs, editing one with URLs copied from the other. 

Just try to get thumbnails posted for now and then work on embedding them at different points within a post.
Remember you can delete your own post if you make a hash of it and a good way to see the syntax of using pics in a post is to Quote that post and examine the layout.
Looks all foreign at first but you soon get the hang of it.

[Doug Brewster]

Thanks for the responses on my issue with inserting pictures; the files were way too large but I think it's going to work this time.

After having this HV multiplier x rayed and taking it apart, in the course of which I've essentially destroyed everything in it, I now know this about it:-

It contained three capacitors
Two of the capacitors had a connection to the ground terminal
There was a 1.55 Megohm series resistor (by measurement and colour code) in the output HT lead connection

I still have a little more excavation to do and hopefully may be able to get a measurement of one of the remaining capacitors. I haven't even seen an identifiable diode; they must have been ground up in the dust.

[/img][/img][/img][/img][/img][/img]

[Gyro]

Glad Resize helped.

If you want to embed the images in-line with text (I noticed your failed '/img's') then you have to click to expand the thumbnail, right-click on it and 'Copy Image Location', then edit your post, inserting the copied image location between the img and /img. (I've omitted the square brackets for obvious reasons)

To be honest, most of the time, it's just not worth the hassle of inserting the full size images in-line. People can just go down to the bottom and click to expand your thumbnails (which still sit there regardless).


P.S. Nice 'teardown', at least it's yielded most of the information that you need. 

[floobydust]

A career in Archaeology looks much easier  Glass-reinforced potting compound?! Take it to a dentist for x-rays.

If two capacitors are connected to GND, then it's not a Cockcroft-Walton configuration. There, an odd number of stages (i.e. 3) has high ripple.

It might be this second configuration, works very well but higher voltage-rated parts are needed.

I'm just guessing. Old Tek was very smart, masters of the art and always a learning experience to dig into.

[james_s]

I've had good luck de-potting things with the liberal application of heat from a hot air gun while carefully scraping away the epoxy while it's soft. Xrays also work well and I can usually do most of the reverse engineering without even cutting into the thing.

Something to remember with HV multipliers if they are driving a CRT is that the CRT itself often forms the final capacitor. So for example the doubler used in some of the B&W vector arcade monitors contains only one capacitor and two diodes.

[Doug Brewster]

A career in Archaeology looks much easier  Glass-reinforced potting compound?! Take it to a dentist for x-rays.

If two capacitors are connected to GND, then it's not a Cockcraft-Walton configuration. There, an odd number of stages (i.e. 3) has high ripple.

It might be this second configuration, works very well but higher voltage-rated parts are needed.

I'm just guessing. Old Tek was very smart, masters of the art and always a learning experience to dig into.

Re the Cockroft Walton observation, yes, it just doesn't fit any CW configuration I can find. The second schematic you've posted fits exactly the - admittedly limited - facts that I have. Does that configuration have a name? Where can I find some more info about it?

[floobydust]

I don't see a name for the other multiplier configurations- this is the third I've seen, from a Korean Hung Change scope.

I wonder how much room you have, as 10kV-20kV disc caps are thick 8mm-12mm
I find a 2,200pF part is about 2,000pF out of the box, which further drops to 1/2 that at rated voltage depending on dielectric.
So it's good to voltage-derate the capacitors as much as practical.
Vishay 615R Series 10kV 15kV caps
Digikey stocks nothing but doorknob caps, but Mouser has some, edit: ouch $7 each

[james_s]

That looks like a Cockroft-Walton multiplier to me, there are several different ways of drawing them and it can look weird if you're not used to the style.

I've bought HV capacitors from Chinese ebay sellers that worked fine. Diodes too.

[floobydust]

Cockcroft-Walton has only one cap to the AC input, the other two have both a cap and diode to the AC input, and only the SC-504 multiplier has two capacitors connected to ground. Each looks different to me and performs differently in Spice.

You can figure the voltages each part will see Vpk, 2*Vpk, 3*vPk to save space and money. I'm lazy and just put in overrated parts, especially if encapsulating.

The eBay/Aliexpress 2CL7x diodes I am not sure about. I was going to measure their real breakdown voltage, as their forward voltage is a lot lower than datasheets.
edi HV diodes spec. 2CL77 20kV Vf=62.5V, 2CL74 14kV Vf=42.5V
I measure Vf=28V on the Aliexpress 2CL77's I bought cheap, so it's a small stack and I thought around 1kV/diode was a ballpark number.
Not sure if the parts are bogus or not.

[james_s]

Isn't the difference just that it's flipped around so that what would normally go to the input voltage is going to ground and vice versa? Since it's AC this shouldn't matter.

https://en.wikipedia.org/wiki/Cockcroft%E2%80%93Walton_generator#/media/File:Cockcroft_Walton_voltage_multiplier_circuit.svg

[Doug Brewster]

Isn't the difference just that it's flipped around so that what would normally go to the input voltage is going to ground and vice versa? Since it's AC this shouldn't matter.

https://en.wikipedia.org/wiki/Cockcroft%E2%80%93Walton_generator#/media/File:Cockcroft_Walton_voltage_multiplier_circuit.svg

In this CW example there's only ever one leg of a single cap to ground...

[Doug Brewster]

I've put together the second of the two possible circuits presented by floobydust in his post #13 above. It does seem to work well as a tripler, perhaps too well, as it's producing just over 50V for 16V pk in. I'm going to call that measurement error. The numbers are actual output 26.9v + 23.8 volts total diode volt drop (8.26 D1. 7.89 D2 and 7.61 D3). This at 32V pkpk in at 50KHz sine. In the pics the left hand meter is rms reading and the scope is set at 5V/div. To answer an observation above about available space I've included a couple of pics to show the scale of things. One pic includes a AA drycell and another shows this multiplier next to one (homemade but the same size) for a 475.

[Doug Brewster]

I should have said in the previous post that the diodes are T73 (12KV I think) and the caps are 2200pF 10KV..

[Doug Brewster]

Well, the pressure's on now. A few minutes after my last post the scope in the images died. F3488 blown; 24V supply to the HT oscillator etc.. Put another fuse in and it ran for a few minutes and blew again. I sure hope that the circuit I'm playing with is going to work at high voltage.

[Doug Brewster]

Excavations done but with some damage to the case as shown in the pictures. It's difficult to work in these multipliers because of the size. The hole that I punched in the case was repaired simply by using some jewellers/watchmakers protective tape to hold the compound in the case during filling and baking.

This multiplier has gone into the scope that failed as mentioned in my previous post in this thread.

The scope is running very nicely with a clean, sharp trace. I've noticed that the trace is taking about 5 secs to appear from power on compared to 2-3 seconds for my other two SC504s. I've used all 10KV caps (because that's what I had) and the last cap is at full output voltage minus diode drops so we'll see how long it lasts. I'm not too concerned if it fails because the modern potting compound I've used is a lot easier to dig out again. Don't ask how I know.