Author Topic: What's inside of an AVTECH AVMP-PS pulse generator?  (Read 351 times)

0 Members and 1 Guest are viewing this topic.

Offline dzseki

  • Frequent Contributor
  • **
  • Posts: 393
  • Country: hu
What's inside of an AVTECH AVMP-PS pulse generator?
« on: October 04, 2019, 07:28:39 pm »
So, at work, in our lab we were looking for something, and at the back of a tall  deep shelf I’ve noted this instrument. The SMA connector with OUT marking catched my attention, otherwise it looked like a small scale/DIY instrument.

After googling up the maker I got the impression they are specialized on fast edge pulse generators, although this unit sure doesn’t looked like a standalone pulse generator. On the front it says AVMP-PS on the backside it says AVMP-3-... AVMP is their pulse generator line indeed, although AVMP-3 should be a GPIB/RS232 controlled unit (this one clearly isn’t that one), and for the -PS I found nothing. I’ve contacted the company but, haven’t heard from them so far. There are various offerings on eBay as well with different AVMP models, with more knobs on the front.
In the mean time I got bored so I popped up the lid to see what’s inside...

Hmm... semirigid coaxial cables and one Pomona magic box plus the power supply. This thing have a bit of DIY feeling in it indeed. At this point I was disappointed because I couldn’t come closer finding out what this really is, only I was more curious...

With some imagination I figured out that I should supply some kind of square wave to the INPUT BNC and that triggers the output pulse with adjustable  amplitude and width. Judging by the simple flat cable that connected the INPUT terminal and the magic box I assumed the repetition rate shouldn't be very high. Nevertheless however hard I tried it did not do anything on its output (terminated with 50 Ohm). All I could see is a few random pulses during power up, but otherwise totally dead.

It wasn't very hard to convince myself that I should open the magic box after the unsuccessfull test. Unfortunately, even after removing all the screw the top plate wasn't budge. After prying it with a heavy duty flat screw driver the cover finally came off only to find that the whole box seemingly flooded with epoxy resin. The thing did not look hopeless though, because there were traces of white silicone RTV along with the hard black epoxy. Also I assumed that if the box has anything to do with RF then it is probably not entirely filled with epoxy.
Again with the help of the heavy duty screw driver (and some luck) I could break through the top epoxy cover  without damaging the internal circuits and finally something interesting started to show up.

About the box, it looks like it is divided into two parts, as far as I can see the part at the back, where power supply goes in is indeed fully filled with epoxy, There should be a DC-DC converter there because there are traces of negative supply voltages inside the box, even though it only has +5V and +24V input.
The front part is the interesting one afterall. It has one wicked construction: The circuit boards after asembly got a conformal coating, then they added the white silicone to fix some parts here and there. After that they put a thin sheet of wood (!) to cover the boards.

then added white silicone to fill in the small gaps between the wood and the box wall, and after that they applied the black epoxy resin to seal the box for sure. Crazy!

Inside it is rather crowded, there are 3 PCBs. All through hole resistors but there are some 1812-ish size SMD caps, there is one TO220 transistor with bent heat fin, and an other one with sawed off heat fin, parts everywhere, lots of baluns and coils. Cables ( and even a series of diodes) run across boards.  Looks like the designer/constructor had a lot of fun, eh?
The cables looked like consistently color coded, so I tried to document all of them where they were tied, as I am attempting to repair/analyze this circuit as much as possible.
Looking at the component date codes, the newest one was from 1988.
The "INPUT" BNC goes right to the first PCB
On the top side there are only passive components.

on the other side there is an MC74F00 quad 2-NAND gate, here comes the INPUT, without any protection let alone termination. further there are some RF transistors hiding, 1pc MMF4049 and 3pcs. 2N5836. There is a coaxial interconnect to the other board.

On this board there is a perpendicular daughterboard with 2 ICs one L272 dual power opamp and an MC1436 high voltage opamp, the function of this board is unclear at the moment.

The third board contains the output stages, this wan't obvious either since not many active components can be seen here from the top. There is a TO-220 (?) device that has the thin long heat fin bent (like old CRT color amplifier transistors), anyway they went for sure and grinded off the part number, just like from the other TO-39 device that is mounted on the underside of the board.

To me it looks like that while this later is a three leg device but only two legs are used. Baluns and coils again, and a tiny small flea under a bunch of huge SMD caps and solder blobs, that very much looks like a tunnel diode, yay!

The output SMA goes directly on the PCB, the other SMA connector that leads to the SMA on the back of the machine labelled generously as "M" (like Monitor?), comes from the same point with an aditional series resitor that looks like a MELF resistor but about twice as long, no color band, DMM thinks it is ~500 Ohm.

I just thought it would be entertaining to some here to see such an oddball instrument.
Later I'll really try to follow the signal flow, and report back what I find...
« Last Edit: October 04, 2019, 07:49:03 pm by dzseki »
HP 1720A scope with HP 1120A probe, EMG 12563 pulse generator, EMG 1257 function generator, MEV TR-1660C bench multimeter
The following users thanked this post: Performa01, radhaz, alm, Fuzzy Star

Offline dzseki

  • Frequent Contributor
  • **
  • Posts: 393
  • Country: hu
Re: What's inside of an AVTECH AVMP-PS pulse generator?
« Reply #1 on: October 08, 2019, 06:37:54 am »
I had no time to advance with this further, but in the meantime I've got a feedback from the manufacturer, and they kindly provided a copy of the original (!) test report of this particular unit.
It seems I was right it needs a TTL input trigger (<1MHz) and the output should be adjustable 0-20V, while the pulse width is adjustable between 5-100ns, the rise and fall times should be 300ps or less.
HP 1720A scope with HP 1120A probe, EMG 12563 pulse generator, EMG 1257 function generator, MEV TR-1660C bench multimeter
The following users thanked this post: Pinkus, radhaz

Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo