Power Transistors for old HP-6920b

[RichardM]

Hi All

I could not resist purchasing this old HP calibrator for <$50 AU

Before turning it on I pulled the series power transistors which are PNP germanium hp part 1850-0168. I think they may be Bendix or HLAB branded according to the manual but after several nights searching I cannot find any specifics about them, somone indicated they were 70v high gain 5A. Two seem ok, one is dead and one looks really leaky. I was looking for advice on a suitable replacement. There looks to be a shortage of TO-3 transistors or perhaps they are fading away naturally. I have some MJ15004 or 2N2955 I could use as replacements ? They are Q8, Q9, Q14 and Q15 in the attached schematic.

Any help appreciated as always.

Regards

Richard

[Kim Christensen]

The best I could find was this info which is pretty close to what you found.

A bit pricey, but they have 2 at Sphere.ca in stock. I have ordered stuff from them about 7 years ago but nothing since but all was OK that time.

[RichardM]

Thanks

I have 2 versions of the service manual and it appears a 2N1545 may be interchangeable for the 1850-0168, at least on the lower pcb board anyway. I will try to search based this transistor also.

[factory]

The change notes seems to suggest they used a matched pair for Q33/34 from s/n prefix 1105A, these are being used as output rectifiers.

Q14/15 are the inverter switching transistors, the inverter may be damaged if the external AC reference signal is lost, the manual does mention this.

David

[Kleinstein]

The Q14 / Q15 part could be a bit tricky as the transistors will also see reverse operation of some kind.

The Q8 and Q9 part looks like a more normal linear operation and chances are good that normal silicon types (e.g. 2N2955) should be OK too.

[RichardM]

Thankyou

The manuals are a little confusing. One manual indicated Q8,9,33,34 are Bendix B1217 (upper PCB) and Q14,15 are Bendix B1203 (lower PCB). The other manual shows all transistors are HLAB 1850-0168. In circuit Q33,34 seem ok so fingers crossed.

Richard

[factory]

You need to check the serial number prefix and use the appropriate manual and/or change sheets.
The 1965 manual has the Bendix parts, the 1967 manual has the HP part numbers, change sheets give the matched pair from change 7 (1105A).

I've had a look in mine (1201A prefix) the matched pair (HP #06920-80093) are Motorola 2N1545, looking at the solder joints these look to be originals, note the error on the circuit diagram, the symbol should be PNP.
The solder on the four 1850-0168 on the top board has been disturbed in the past, I have three Motorola 1850-0168 and one TI 1850-0168 fitted to mine.

I would suggest checking the larger capacitors, some of mine are showing high ESR and/or cracked seals and could do with replacing. I seem to remember from another thread, someone mentioning the voltage rating was a bit low on one of the capacitors.

Edit: more errors in the later manual, it lists C31 & C32 as ceramic caps  , older manual is correct with electrolytics listed.

David

[RichardM]

I replaced all 4 of these transistors with MJ15004, rewired a new IEC power socket and turned it on with my dimb bulb. Amazingly it works for DC volts and amps. Just need to test AC now. I purchased all new axial capacitors so I will systematically replace those now. I have noticed the ten turn pot appears to need a clean. Can this be opened up and cleaned ?

Richard

[coppercone2]

post picture of potentiometer, I have not seen a HP with a ten turn on it that is unsealed IIRC. Unless there are screws on the back you are SOL I think. I have thought about cutting them open and gluing them back together with a fine saw, but that is just conjecture. If the housing is thick enough you might be able to press fit pins in there with a bit of glue and that might hold for a good while, or bend something out of sheet metal that wraps around it, like the part used to gang potentiometers, or like wafer switches (that is a sandwich between two plates, spacers with screws). I have a bad ten turn that I might cut open, but I suspect it might be difficult to reassemble.

Would be nice to figure out if this is feasible actually, because they kinda add alot of price to a project compared to say two screws, some brass tubing and sheet metal.

I did replace the germanium transistor on some HP supplies, it all seemed to work OK.

I looked at my spare ten turn pots, one of them has three screws on the back (spectrol),the ducan brand one has 4 melted plastic studs (possibly drillable for nice disassembly), and the 4 turn bourns (that turns without hiting a stop forever) looks like the back was just glued on some how, but I suspect this has some kind of intend in it, that is if you cut it off, you would need to either fabricate a new part or add a spacer to glue the old one back on, like I mean I think it has seat for the shaft on the back maybe. But even so, you could make a mold of this part out of like JB weld and put a ring on top then file or machine it down to the correct height if you cut it flat.

[alm]

post picture of potentiometer, I have not seen a HP with a ten turn on it that is unsealed IIRC. Unless there are screws on the back you are SOL I think. I have thought about cutting them open and gluing them back together with a fine saw, but that is just conjecture.

I've seen ten tun pots that had a metal clip you could bend to open (Bourns), and some (not in HP equipment) that had the front part with the shaft press fit in the plastic shell, and could be separated by inserting a sharp object between the front and the shell.

[RichardM]

I finally got back to the HP-6920b today. DCV and DCA output is fine but I have no real output in AC mode and the output bulb does not illuminate. Back to my hobbyist troubleshooting skills. What circuit or components are key to the AC output ? The 1V AC reference seems fine.

Regards

Richard

[factory]

Difficult to say, switch 1 middle position (AC) is connected to many places, the demodulator seems only to be used in AC (bypassed in DC) though.

David

[RichardM]

Thanks David. Yes I find the schematic and use of so many transformers hard to follow. Also knowing where circuit ground is. I generally start to check diodes and transistors. Are Q33/34 used for both AC and DC ?  I thought so.

[factory]

No Q33/34 are only used as rectifiers for the DC mode, again note the middle position of switch S1 bypasses these transistors, lower half of the output transformer winding is left open circuit.

It would have been nice if the Harrison division had added extra diagrams for the AC & DC modes, with the relevant parts highlighted, similar to the flow diagrams found in some other manuals.

The ground reference for measurement looks to be -S on the diagram, this appears to be the chassis ground and in the first four positions of the range switch S2 is also connected to the - output jack.

David

[RichardM]

I really don't know what I should be observing from the demodulator circuit ? DCV and DCA working fine. All small transistors check ok in circuit. I do see a square wave at collectors of Q14 and Q15 but not 10V amplitude that is shown in one version of the service manual. Love to know what another owner of a 6920b may observe. Maybe the replacement MJ15004 transistors are not suitable for the AC ? Although I see 1V AC reference is fine.

Cheers

[RichardM]

Another observation. The AC output is completely different (still incorrect) depending if I power the 6920b on (a) 115V directly from the variac or (b) through my dimb bulb tester (buld not lit) which is powered from the same variac ?

[coppercone2]

did you try to load the output, it says it does 5 amps on AC.

[RichardM]

No I haven’t. What would loading the AC output show ?

[coppercone2]

not sure, but if changing the input impedance does something maybe the output impedance might do something more obvious

[coppercone2]

I wired one of those germanium transistors backwards in something HP (the one that have wire soldered to the lead) and it was doing funky stuff related to power levels IIRC, when the output got over 10mA or so.

[factory]

I really don't know what I should be observing from the demodulator circuit ? DCV and DCA working fine. All small transistors check ok in circuit. I do see a square wave at collectors of Q14 and Q15 but not 10V amplitude that is shown in one version of the service manual. Love to know what another owner of a 6920b may observe. Maybe the replacement MJ15004 transistors are not suitable for the AC ? Although I see 1V AC reference is fine.

Cheers

Yes I could check waveforms at Q14/15 (dependant on spare time at weekend), any suggestion of front panel settings? to have them set the same for comparison.
I'm sure the AC output is a crude sinewave, as it's referenced to the mains waveform. 

Another observation. The AC output is completely different (still incorrect) depending if I power the 6920b on (a) 115V directly from the variac or (b) through my dimb bulb tester (buld not lit) which is powered from the same variac ?

How many volts are being dropped across the bulb? even if not visibly lit it will still drop voltage, hopefully not too much though.

David

[RichardM]

Thanks David

I will compile a table with measurements and if you were able to check just a few against your unit that would be hugely appreciated.

I get a nice square wave on the collectors of Q14/15 in DC mode but the amplitude is less than half the 10V outlined in the service manual. I did change Q14/15 to MJ15004 transistors as one was dead. If I then switch to AC mode then I see a distorted square wave with the same amplitude that looks nothing like the sine wave in the manual. The amplitude at Q28/29 is also down in amplitude.

My unit has the 005 option for 50hz mains but is still wired for 115V. 50Hz is compatible with Australia but we have 230V mains hence why I am using a variac. When the AC output is ON and I adjust the variac output by as little as 1VAC I see an immediate and sometimes larger change at the output ?

Richard

[RichardM]

Some DC readings to begin with

If I set the output dial to 1000, Voltage range 1V, DC selected I get the following

(1) Q14/15 Collector, Output=off, square wave -0.5V to +0.5V or 1V p.p 100Hz
(2) Q14/15 Collector, Output=on, square wave -5V to +5V or 10V p.p 100Hz
(3) Q30/31 Collector, Output=off, sine type wave +2.5V to +7.5V or 5V p.p 50Hz
(4) Q30/31 Collector, Output=on, sine type wave +7V to +12V or 5V p.p 50Hz
(5) Q28/29 Collector, Output=on, square wave +4.5V to +9.5V or 5V p.p 100Hz

I have also noticed after a touch test that Q21 gets scorching hot. It is a metal can but Q19/20 are plastic transistor TO-92 style. Should it be getting that hot ? Bit suspicious ?

Richard

[Kleinstein]

IN the old days they did run some transistors quite hot - at least silicon transistors, not so much with germanium types.
If you can still touch them without burning the finger it is probybly still OK.

It is hard to tell from the schematics fragment how much power Q21 would normally get, but it could be quite a bit as the voltge ratings of the capacitors look high. 

[RichardM]

Ok, thanks. It is hot enough to burn if you held your finger on it. It's in the AC reference circuit and I can see the 1VAC output fine so I guess that Q21 must be OK. Does anyone know what transistor could replace a 4JX16B533 ? Cannot seem to find a datasheet for this NPN which is supposed to be for Q19,20 and 21. Q19 and 20 are plastic packages in my unit.