Philips PM6614 Frequency Counter

[Philips210]

Hello everyone.

Today, I discovered a Philips PM6614 frequency counter that I've had for many years but never got around to looking at it. It's a Swedish built unit that measures up to 520MHz.
The counter is in reasonable condition. Although the top and bottom covers look poor, they aren't damaged, so should respond to a new paint job.
Opening the unit, I noticed the mains transformer is missing. This is a low profile type and probably specific to this range of frequency counters.
On the rear panel, there's an option for an external DC supply which is handy.

The build quality looks good which is what you'd expect from Philips test equipment. The counter appears to be from around 1974 as most of the ICs have date codes for that year.
I notice what looks like a number of solid aluminium electrolytic capacitors, the orange coloured components on the main board. Am I correct in assuming this and would it be safe to leave them in place rather than replace them?

I assume there are a few specific unobtanium ICs in there too?
I haven't tested the unit as yet but wondered if anyone with experience of these counters can make any suggestions as to what to look out for?

Also, if anyone has a scrapped unit with a good mains transformer they want to sell I'd be interested. Please see a few attached pics.

Regards,
Symon

[WME_bill]

PM6614.
I think you can be encouraged for these mainly use commodity integrated circuits and components. I have one sittting in front of me now, with the manual. Nice machine. Dates from 1975 but went on for years.
The transformer is a straight-forward 20-24V, and I have measured the current load as <270ma. So a 24V toroidal 15Va should be ample. Feeds into a switching mode power supply.
The battery supply is >18V.
The basic counter to 80Mhz (A channel) uses transistors and highspeed TTL.
The 520Mhz channel B may be more difficult. It uses BA379 PIN diodes and two hybrid UHF amplifiers Mullard OM185. The Mullard OM335 is similar. The alternatives for the Mullard hybrid family OM335 - OM345 have been discussed here on the Forum from 2010 for the Marconi signal generators TF2022.
It also uses Philips GZF1201 locmos decade counters. I am not sure if still available, but they do seem fairly reliable.
The other problem may be the Plessey divider SP8670D (600Mhz,x8). I think Plessey went into GEC and when that collapsed, these counter ICs went to Siemens. I am sure similar is around somewhere.

I assume you have the manual.
wme_Bill m0wpn

[Philips210]

Hello Bill.

That's very helpful information. The toroidal mains transformer is a good solution to the problem.
I didn't realise at first but this counter can also be used as a timer which is very useful.
I haven't got any service information on the PM6614 but if you have a copy of the manual that would be great. Thanks very much.

Regards,
Symon

[orbanp1]

Hi Symon,

It looks like you counter-timer is part of the PM-661x family.
The PM-6615 is the "1GHz member" of the family.
Chances are that the PM-6614 is a similar machine, with a different prescaler.
The schematics for the 6615 can be found on elektrotanya. That might be helpful.
Granted, this is all speculation...

Regards, Peter

[Philips210]

Hello Peter.

Thanks for the link, that's very useful and hopefully most of it will be applicable to the 6614. Elektrotanya often provides useful service data, a very handy site.

Best regards,
Symon

[Philips210]

Just to say thanks to Bill (WME_bill) who kindly helped me with comprehensive service information for my counter. This should enable me to tackle any servicing and repairs with more confidence.

I will study the circuits and assess what needs checking before applying power. If I can get the counter working with an external supply, then I'll purchase a small toroidal mains transformer and devise a means of fitting it to the main PCB without having to drill additional holes, and thus spoil the board. I thought about making a special plate so that it can be fitted using the existing four mounting holes.

I will post an update in due course.

Regards,
Symon

[Philips210]

Hello again.

I've done some tests on the frequency counter and found a few things needing attention. As pointed out earlier, the mains transformer is missing. On the inside of the top panel I noticed a small burn mark where something molten has hit the panel, possibly a small blast from the original transformer. I also noticed the LT fuse fitted was a T3.15A where it should have been a F1.6A.

I originally had the idea that the BY164 bridge rectifier may have gone s/c and ruined the mains transformer as there's no fuse in between to protect the transformer. Checking the bridge I was getting an o/c reading on one of the diodes. I decided to remove it for testing/replacement which is easier said than done as the PCB has through plated holes making it difficult to extract components. Controlled brute force is the only answer to avoid damaging the PCB. I worked the bridge back and forth about 100 times to snap it off the board. I then applied fresh solder to each of the four pins and then carefully extracted them. Desoldering braid then sucked the remaining solder away leaving clean undamaged pads.

When I checked the bridge it was not o/c as originally thought although the volt drop was markedly different on the suspect diode compared to the other three so it needed replacing anyway. The original BY164 was replaced with an RS Components 261-491 200V 1.4A bridge, the pinout being the same as the BY164.
I also found R1124 (1 Ohm) that feeds the bridge rectifier's o/p to the reservoir capacitor C152 (680uF) was o/c. I removed it and it look perfect but was certainly open.

So, I am left wondering what originally happened. I wondered if there was a largish mains voltage surge that wrecked the transformer and caused the other damage. Anyway, I replaced all the faulty parts and did some further tests. Cold resistance checks across the supply rails didn't show any problems so I decided to give it a test. I used the external battery supply connections on the back of the unit which were connected to my bench power supply. I gingerly appled 12V, and well, it came to life but the display didn't look good.
The immediate issue with the display was the 'b' segments weren't lighting on any of the nine display sections. Also the remaining lit segments weren't that bright and seemed a little erratic. The top 'a' segments appeared a little brighter than the others. I slowly increased the output from my bench supply up to 20V whereby the segments were evenly bright.

So, the real problem is with the 'b' segments. I think it's either the display itself or the HV driver chip IC181 (DM8884A). The display is of Panaplex type which uses Nixie technology with neon gas but the digits are like normal seven segment displays. Each segment being a cathode. The next step will be to try and test the display to confirm if all the segments will light.

I'm kind of disappointed as it looks like it may be uneconomic to repair this counter which is a shame but I won't give in yet.

Please see a few pics. In the first, the burn mark on the top panel. In the second, the bridge rectifier, R1124 and the LT fuse have been replaced. Pic 3 shows the defective display. The display is now slightly brighter since it's been on for a while. I haven't as yet done any frequency measurements but the timer section appears OK as the gating LED flashes at the correct intervals.

More to follow...

Regards,
Symon

[factory]

Hopefully a bad connection somewhere and not a dead output on the display decoder/driver IC, but those National Panaplex driver ICs do have a bad reputation for outputs failing (causing dead segments), the display is probably a Panaplex II, another brand name used by Burroughs.

Looks like the DM8884A was/or might still be made by Dionics as DI-8884A, but none advertised for sale at the moment.

Depending on space available, it could be possible to build a replacement using regular logic ICs, with a seven segment decoder, some higher voltage driver transistors on the outputs and resistors to limit the current (instead of the individual current source used). Take note of the drive current set by the external resistor in the counter, too high current and the display won't last long.

As for the transformer failure I notice the PM6615 has a thyristor based crowbar circuit (typical of Philips scopes), this would short the supply to blow the fuse if the 5V supply was too high, the higher value fuse was probably a very bad idea.
I'm surprised no damage was visible on the resistor, kind of would expect a burn mark if the capacitor went short, as Philips electrolytics sometimes do.

Those orange solid aluminium caps have a much better reputation.

David

[Philips210]

Hello David.

Thanks for your helpful remarks, it's reassuring to know about the solid aluminium electros reliability.
Yes, I was quite surprised to see no damage on R1124. I reckoned C152 could have caused the failure. Maybe the unit was long dormant and the PSU electros needed reforming. The initial excess leakage current in C152 could have resulted in the o/c R1152. C152 seems fine now, no warming but it wouldn't be a bad idea to properly check the leakage current. The problem is the extensive use of through hole plated PCB pads which makes servicing not much fun. Normally, I'd remove all the electros for thorough testing/reforming.

Looking at the back of the display it seems to have 18 pins but on the component overlay and circuit diagrams it shows 19 pins. I'd prefer the driver IC to be faulty rather than the display but a connection issue may be the cause. It looks a little tricky to get to the front of the display to dismantle it and check.

Another thought, each segment is fed from the driver IC via a 2k2 resistor. Maybe swap over say the 'a' segment feed with the faulty 'b' segment feed. If the dead segments are now the 'a' ones then the driver IC is to blame. Does that seem reasonable?

I like your suggestion of building a new circuit to replace the driver IC.

Any further help is welcome. Thanks.

Regards,
Symon

EDIT: One thing to be aware of when working on these counters is the high voltage supplies to the display circuit seem to hang around for some considerable time after switching off. There are no bleed resistors in this part of the circuit, so proceed with caution and discharge the supplies via a resistor of say 4.7kOhms and recheck. The HV supplies smoothing caps seem to recharge so be wary of this.

[factory]

If only there was a datasheet to confirm the pinout, as the free pdf isn't terribly clear.
It's possible there are unused connections on the display, either N/C and/or maybe an extra connection for the comma/colon symbol.

Yes swapping resistor connections should help confirm if the display or decoder/driver is at fault.

I wonder if it would also be possible to use the other Panaplex decoder IC, the DM8880, this only has the seven segment driver outputs, the decimal point would need a separate driver circuit & current limiting. The outputs for BCD 10 & above do appear to be different though. The DM8880 is also difficult to obtain these days too.

David

[Philips210]

I've just realised there are a few differences between my counter and the one depicted in the service information. For a start, on my counter there isn't a 470k SIL resistor network connected to the decoder/driver IC outputs. Also the 2k2 current limiting resistors that connect to the display segments are layed out differently. As I mentioned earlier, the display pinning looks different too.
Please see attached pic. It would appear to be an early version with many 1974 date codes on the ICs. In a user only manual for the PM661x series, I noticed a pic of the main board which appears to be the same as mine. It seems strange that Philips didn't mention the two versions in their service information.

I checked with an Ohmmeter from pins 16 and 17 of the decoder/driver IC, segments 'b' and 'a' respectively, and the 2k2 resistors associated with those pins are the ones at the top right of the decoder/driver IC in the attached pic. Tomorrow, I'll swap over the connections from these resistors to the display. Hopefully it will at least exonerate the display.

Regards,
Symon

[Philips210]

Hello again.

I swapped over the segment 'a' and 'b' outputs from the DM8884A decoder/driver IC and now the dead segments are the 'a' ones, so the display is thankfully OK.

Before I did the swap, I tested the counter and found the display looked weak again, being dim and unstable in many segments. I noticed there were no decimal points either. The missing 'b' segments were the same.

I also lifted all eight 2k2 resistors and connected the known good 'a' output from the driver IC to each segment on the display. Each group of segments including the decimal points lit correctly proving the display is good.

Interestingly, if I tried output 'c' on any segment, there was basically a blank display. This also seemed to be the case trying other outputs to various segments. I thought that I had done further damage to the IC but when I reconnected the 2k2 resistors, the display returned to its previous state. This was probably due to the blanking signal which is linked to all nine segment output stages in the driver IC. I expect the output stages are blanked if several outputs are disconnected from the display cathodes.

Incidentally, I noticed segments 'a' are generally brighter than the others but when tested individually they appeared to be the same brightness as the others. There is clearly a problem with the DM8884A decoder/driver IC.

Finding a replacement is going to be difficult and no doubt expensive. I wouldn't mind trying to build a replacement, using say a SN7447 BCD to seven segment display decoder/driver IC, and suitable transistor output stages, to drive the Panaflex tube cathodes. A separate output stage will also be required for the decimal point. The blanking signal will have to be included in the new circuit so adding further complication. So far, I haven't been able to find a standard type of circuit to do this. I will have a think about the best way to proceed.

Any further suggestions and help would be great. Thanks.


Regards,
Symon

[factory]

How is the blanking being done in the PM6614, is it something to do with the current programming input?
Both the DM8884A & 74LS47 will blank a digit with BCD 15, but maybe isn't how Philips used it.

David

[Philips210]

David, yes the blanking input is applied to pin 1 of the IC.

A copy of the PM6615 circuit which is similar to the PM6614 can be found here:

https://elektrotanya.com/philips_pm6.../download.html

So using the 7447 may not be too difficult after all.

Regards,
Symon.

[Dickie]

FYI, there is an old Marconi Instruments spec. ana. , TF2370 that used a Panaplex display driven by a 7-segment decoder for its counter display. The manual is available from the BAMA archive: https://bama.edebris.com/manuals/marconi/tf2370/

It will take a bit of reverse engineering to work out how it does it, but it might give you a few ideas.

[Philips210]

Thanks Richard, that's very good. I'll have a look at the circuits which look very applicable.

Regards,
Symon

[factory]

Originally Posted by Philips210

David, yes the blanking input is applied to pin 1 of the IC.

A copy of the PM6615 circuit which is similar to the PM6614 can be found here:

https://elektrotanya.com/philips_pm6.../download.html

So using the 7447 may not be too difficult after all.

Regards,
Symon.

Now I've found a readable DM8884A datasheet it does show a 7405 inverter IC being suggested for grounding the current programming input to blank the display.
DM8884A.pdf

I've been using those PM6615 circuits, it doesn't seem to ground the programming input that I can see, seems to be something related to the clock signal & battery low, associated with both the current programming supply & digit select transistors.

Wonder if the SN75468 (TTL input version) 7x transistor array could be used with a resistor network for the current source function. I found the SN75468 by accident mentioned in a ULN transistor array datasheet, as a higher voltage version, gas discharge displays are one of the suggested applications. Of course individual HV transistors could also be used.

David

[Philips210]

David,

there's a circuit description in the attached user manual. The blanking is a little complicated.

Regards,
Symon

[factory]

Originally Posted by Dickie

FYI, there is an old Marconi Instruments spec. ana. , TF2370 that used a Panaplex display driven by a 7-segment decoder for its counter display. The manual is available from the BAMA archive: https://bama.edebris.com/manuals/marconi/tf2370/

It will take a bit of reverse engineering to work out how it does it, but it might give you a few ideas.

Looks like they choose the 74LS48 (now obsolete along with 74LS49), same pinout as the 7447, but the 7448 has internal pull up resistors on the outputs, 7447 is open collector.

David

[factory]

Originally Posted by Philips210

David,

there's a circuit description in the attached user manual. The blanking is a little complicated.

Regards,
Symon

It certainly is, I note the mention of two versions of the circuit as well (maybe why your counter doesn't have the 470k SIL resistor network), later ones blank via the digit select too.
Both the 7447 & 7448 do have a blanking input, I guess you will need to experiment to get it to work and prevent the unwanted glowing from the display segments that are off.

And the operation section, mentions a low battery indicator, that blinks the display when the internal batteries (an optional extra) are low.

David