EIP 545A 18GHz counter questions. Signature analysers?

[Chris Wilson]

I have acquired for very little money an EIP545A 18GHz counter. It was acquired displaying just dashes. I checked the voltages and found the PSU outputting well away from the manual figures. I corrected these and checked supplies for ripple and they were all good. I then decided to remove any board that was not essential to operation. It came with a GPIB option board, so I pulled it. Luckily this did some good and a display appeared. There were three caps on this board, as a matter of interest I pulled a leg on each and tested them, all were OK. Refitting it killed the display back to "dashes" again, so I set it aside as "unknown fault" and continued. I can get it to accurately display up to EXACTLY 188.9999 MHz. Inputting 190.0000 MHz gives zeros. It has three frequency bands. Band one works perfectly right up and well beyond its 10Hz to 100MHz range. Band two often doesn't work at all, it's a 50 ohm input, compared to the 1meg 20pF Band 1 input. Sometimes by ramping the frequency up from 100 MHz I can get a reading of up to 189.9999 MHz. Band 3, the GHz range, doesn't work at all. Only zeros are displayed.

The machine has quite a good range of self diagnosis tests. Tests are via the key pad. The first test checks the VCO and other stuff, and should display an accurate 200.0000 MHz. It displays well over, and the display isn't stable, it hovers over several KHz. There's a tree menu to see where this issue could lie. One limb suggests using a known good counter on the output of the VCO. I did this and the output is miles high in frequency, about double, and unstable. The tree menus goes on to suggest a phase lock circuitry fault.

Further tests seem to depend on using something called a Signature Analyser, which I have never even heard of They suggest an HP5004A, which is apparently pretty ancient. Is there a cheap way of acquiring something to do this sort of testing, or is it goodnight Vienna?

It's a nice old thing, and I would quite like a means of counting into the higher frequencies it offers, but don't want to spend too much time or money on it. It works a damned sight more than when it first landed, which has kind of given me incentive to push a bit further, given I have a .pdf copy of the repair manual. Thanks!

[G0HZU_JMR]

Hi Chris

Edit:
I've removed my first reply about the 5004 as I think it's a bit more complicated than just being a basic data decoder.

It looks like it compresses data into a form of 4 digit checksum or signature so you can compare it with data being sent in a good counter.

[G0HZU_JMR]

I don't know if it's possible to emulate the 5400A algorithm in a PIC/AVR chip but presumably it does some form of high speed CRC on a stream of data and spits out a 16bit key at the end that it displays on its screen.

Somebody somewhere has probably emulated it on a different platform?

[PETERg0rsq]

Chris

I have a 5004 signiture analyser.

I would be willing to lend it to you, but it would need to be face to face with a deposit to the value of replacing it!

Of course when it is returned, you would get it back

I am in St Helens so if you can collect it and return it and are interested in borrowing it, let me know.

Peter

[G0HZU_JMR]

I just downloaded the service manual and this is old school Motorola stuff. eg 6809 processor.

The block diagram shows it uses classic old Motorola PIAs to interface to each section. This makes it easier to faultfind.

If the band 1 counter works then I'd assume the main processor + RAM and PROM and address decoding etc are healthy. Same goes for the heart of the counter and the display.


So something is wrong with the other ranges that use mixing to convert the signal to appear within the range of the basic 125MHz counter. So you could check the downconversion circuit see if it is operating with the correct LO frequency across the range.

I'm out of time tonight but you should be able to monitor the status of the downconverter on range 2 to see if it is converting the signal to the correct target freq within 0-125MHz.

Your HP8568B would be a great diagnostic tool to use here

On range 3 the YIG is controlled via a PIA port so it should be possible to see this steering the YIG via the DAC.

Presumably you can manually steer the VCO and YIG in some way via special modes using the keypad and the service manual? I suspect that once you get range 2 working OK then there's a good chance that range 3 will suddenly start to behave as well.

[Chris Wilson]

Great reply as always Jeremy, but I sometimes feel you have a misguided opinion of my knowledge level I *THINK* I follow some of that, but whether to the degree needed to safely diagnose further I am not really sure!

The coward's way out is to diagnose to board level and buy a £43 board from China from someone on Ebay breaking one. But I would like to see if I can learn from this, so will make time to digest the manual further and see whether I can progress to a higher level of diagnosis.

You do realize you have left yourself open to more e-mails form me?

Thanks Jeremy, you really are a most helpful and knowledgeable chap, I feel positively Plebian.

[Chris Wilson]

Originally Posted by PETERg0rsq

Chris

I have a 5004 signiture analyser.

I would be willing to lend it to you, but it would need to be face to face with a deposit to the value of replacing it!

Of course when it is returned, you would get it back

I am in St Helens so if you can collect it and return it and are interested in borrowing it, let me know.

Peter

Hi Peter, that's very kind of you. I quite fancy getting one for myself and have an eye on Ebay. If I can proceed no further without one of these I'll get back to you, you aren't that far away and a deposit is no problem, I quite understand. I still have a few items of engine test gear I have lent that still haven't made their way back home, it's very annoying.

[dinkydi]

That's an exceptional counter you are working on - far better than you would normally come across. Take things slowly, stick with it and I am sure you will be rewarded. It's perseverance that will get you there, not knowledge or even brilliance. (Though I am sure you have all those qualities!)

You probably don't need a signature analyzer (most equipment can be repaired before you get to that stage). However, by all means get one to play with. They usually go very cheaply at auction because nobody has any interest in them (even if they know what they do!)

Unless you find a very, very cheap HP 5004a, I would wait for the more common, and improved, HP 5006a to turn up. This should also be inexpensive.

Peter

[Chris Wilson]

Thanks, would quite like a signature analyser, I have a thing about collecting odd bits of stuff, as my wife will tell you

Yes, it's a nice old thing, not cosmetically perfect, but OK and I have seen far worse. I'll stick with it. Just in case it has any significance, TEST 01 which should display 200MHz exactly, actually always shows 253.5**** the last four digits are constantly changing. I was wondering if the main offset was significant from 200MHz?

Thanks Peter.

[dinkydi]

Off the top of my head, I think you need to identify the down-converter circuit (phase-lock loop), and find its input and output.

Look at the input and output signals with equipment at your disposal (high input impedance oscilloscope/counter etc), check amplitude (if possible) and measure frequency. See that the input/output frequency ratio is the correct integer for Range 2 inputs of, say, 100MHz and also 250MHz.

Depending on your results it should be obvious what to do next.

Peter

[Chris Wilson]

Originally Posted by dinkydi

Off the top of my head, I think you need to identify the down-converter circuit (phase-lock loop), and find its input and output.

Look at the input and output signals with equipment at your disposal (high input impedance oscilloscope/counter etc), check amplitude (if possible) and measure frequency. See that the input/output frequency ratio is the correct integer for Range 2 inputs of, say, 100MHz and also 250MHz.

Depending on your results it should be obvious what to do next.

Peter

Thanks Peter, I believe this is the relevant schematic, I'll pluck up courage to have a poke about, specific instruction as to where to poke appreciated, if anyone has the time and inclination! Problem I have is uploading high enough resolution schematics as to be useful given the forum file size upper limit. I don't think I am allowed to link to the file on my web site, either, am I? Can I link to an ftp site? Cheers.

EDIT, can't see anything in the forum rules about linking to a web site, so a clearer schematic is at http://www.gatesgarth.com/phaselockloop.jpg

[Chris Wilson]

U6 is missing off the board. A Chinese seller has a machine he is breaking, the photos show this socket is populated on his board....

[woodchips]

The signature analyser just records a serial data stream and gives the result as a CRC, cyclic redundancy check, just like the error detection on disc drives. A shift register with feedback and exclusive OR gates, the feedback is the polynomial, x15+x7+x1 or something, can't find it at the moment.

The Tektronix 308 data analyser has a signature analyser function, as well as parallel and serial logic analyser. Sure I have a spare one, no idea if it works, yours for the postage if you want it. It is in my spares pile, nice little CRT but have more than one anyway. Don't have any parallel data pods, reason they never got sold.

Bob

[Radio Wrangler]

If you do a search for the Hewlett-Packard Journal archive, and look up the signature analyser, you should find quite a detailed explanation of the thing, if I remember rightly.

David

[G0HZU_JMR]

Whilst I suspect the fault(s) may well be something simple like a dirty card edge connector somewhere I did skim through the manual today and this counter should be repairable even if it has multiple faults. Obviously there will be a problem if something like the YIG has failed but because the basic range 1 works then this proves that most of the counter is healthy.

I can talk you through how ( I think) the range 2 synthesiser board works and how to probe it when doing the test 2 (200MHz) self test if you like? This is best done with a tutorial video...

But here's some basic info anyway:

It looks like a very complicated schematic at first but it is really just a very basic synthesiser (old school discrete stuff again)

i.e. a single loop PLL using a 40/41 dual modulus prescaler and some discrete logic for the divide by N. The N code gets sent from the processor to the '192 and '193 divider chips via the 6821 PIA via its port A and port B pins. The phase detector runs at 50kHz.

This makes it very easy to faultfind in terms of correct programming for N and signal tracing for the 200MHz self test (i.e. VCO would be at 400MHz).

So set it to run the 200MHz self test again and faultfind as below:

During this test you could either probe for the correct N code on the PIA pins on A108 with a DVM or see if the signal from the VCO has been divided by 8000 before it reaches the phase detector at TP10. The latter is easier and a more thorough test of the divider chain.

400MHz divided by 8000 = 50kHz

But if you do want to check the N code programming from the PIA then I 'think' you should only see a logic 1 on 'port B D5' of U7 (this is pin 15 of U7 on A108)

All the other pins on port A and port B should be logic 0. i.e. all pins 2 through 17 on U7 PIA MC6821 should be logic 0 apart from pin 15 for the correct code for 400MHz at the VCO (but I might be wrong...)


Alternatively, is the freq at TP10 in your attached schematic exactly 1/8000 the (unlocked?) UHF frequency arriving at the J1 coax input on A108 module on the left edge of the schematic?

If it was locked and working these should be 400MHz at J1 and 50kHz at TP10 (factor of 8000 difference) but as they are unlocked they could be 450MHz and 56.25kHz (same factor of 8000 difference)

So the test here is to see if the processor has successfully programmed the PIA (U7) and the PIA has preset the N counter chips and the whole divider chain is dividing by 8000 for the 200MHz self test routine.

If you don't have another counter for this test you can probe both TP10 and the J1 points using a x10 scope probe connected to your HP8568B on the big N connector input (AC coupled). Don't use a x1 scope probe and don't be tempted to use the BNC input connector on the 8568B that is DC coupled as there is a risk of damage to the analyser. Zoom in the span quite close so you can measure each one accurately with the marker function although the J1 signal will probably look quite jittery and will wander around a bit.

Whilst you are probing TP10 have a look at the nearby pin 1 of the phase detector chip U14 (MC4044) This should be a 50kHz square wave on a scope and this signal is the 10MHz reference divided by 200 so should always be there and should be a rock solid 50kHz square wave on a scope.

If these tests pass then I guess it's good news and the fault may be somewhere in the loop filter and error signal that steers the VCO.

I should point out that I've never been inside one of these counters and all the above is based on a quick skim through the manual so I may have got some things wrong...

[Chris Wilson]

Blimey, that's a comprehensive post Jeremy, thanks!!! Before I get carried away with more testing, did you spot my post with regard to U6 being missing? As far as I can figure out how this thing ticks, this is a vital component, and it can never have worked properly if that was missing. I have ordered a new one , it's a MC10131L and I found one from Little Diode who I have used before. I have also got screen shots of the pdf pages showing the text of how this area is supposed to work, but you seem to have good access to a copy of the manual on your own PC. I can't see how U6 on board A108 can be just for some option, it looks essential.

This A108 board also has two multi tun pots on it, right at the back end, on the top. They are designated R6 and R10, and according to the manual are listed as "not used". Whatever, someone has fiddled with them as the paint seals are broken and the screw slots a bit ragged. They are soldered into tracks that radiate elsewhere, so not sure if they really are unused, and why they are actually there.The Chinese vendor on the bay shows this baord and his has these two multi turn pots on his, as well.

I do have a working counter, a Racal Dana 9908, it only goes to 1GHz, but is pretty simple, and is a timer as well as a counter. I am about to watch a You Tube video by Alan, W2AEW, from whose videos on oscilloscope training I have learnt a lot. It's on basic phase locked loops.

I am still very wary of using the 8568B in and around circuits I am not 100% conversant with, in case I do it a mischief. Using my scope is far less stressful

Let me know what you think about U6 being absent without leave please. I suspect until the socket is populated it won't work... I love how you say it's dead basic, I am sure it is compared to the 8568, but it looks pretty darned complex to me Thanks Jeremy, and all who have given advice and offered assistance, much appreciated.

[G0HZU_JMR]

Hi Chris
Yes, sorry I missed the bit about U6 being missing.
That does look to be part of the 40/41 DM prescaler and presumably it uses its two flip flops to get 40/41 from the 10/11 prescaler U5.

Note that on a healthy example of this synthesiser you don't 'need' dual modulus control for synthesising 400MHz as 50kHz x 40 is divisible from 400MHz by a fixed code of 200.

But I can't immediately see how U6 can be omitted as it needs to divide by 40/41 and not 10/11

[Radio Wrangler]

Chris,

When the microprocessor controlled microwave counters came out, they had a neat trick. Like the heterodyne oscillator instruments before them, they had extended coverage using harmonics of their LO, so the manufacturers used a synthesiser and had a facility to offset its frequency a known, small amount. With an incoming signal at X GHz, they would tune the synth around until the 200MHz counter on the IF gave a stable reading in its happy zone. It then offset the synth a known amount, and measured the change in the IF count. By comparing, it could determine which order harmonic was mixing the signal, and the direction of shift told whether it was a USB or LSB mixing product.

So while working on it, you may see the synth frequency being moved around.

Quite clever boxes.

David

[Chris Wilson]

Thanks David. As far as I can tell, and from what Jeremy has said above, I believe U6 has been removed for some reason. Hopefully a new one will arrive tomorrow, so I will install it and see what happens next. I know Jeremy is a closet movie magnate and loves getting on camera, so I may take him up on another tutorial video His bespoke YouTube one on checking out HP8568 and family spectrum analysers was excellent, and gave me the confidence to buy my elderly 8568.

So, Jeremy, lights.... camera..... ACTION

(Just kidding, put the make up away...)


I have also found that there is a dedicated EIP reflector on the Yahoo Technical Groups, to which I have sent a membership request.

[G0HZU_JMR]

LOL
I don't really do presentations that often because I have a poor speaking/presentation voice. Most people tend to when I give a presentation


Hopefully the counter will spring to life on range 2 with the U6 chip reinstalled but it would be worth checking there are no signs of bodgery on the back of its connections because maybe there could be shorts or links fitted.

I can't see how a 10/11 DM prescaler (U5) can be used without the two flip flop sections in U6 as the VCO freq divided by only 10 would be a bit high for the LS192 logic chips to reliably clock.

Plus the dual modulus control circuit (U11 etc) will be discretely designed around supporting DM division ratios of 40/41 and the code in the PROMs on the processor board will be coded up to support the larger DM division ratio of 40/41 ahead of the N divider chips.