B&K 700 Tube Tester plate voltage too high?

[agent_power]

Hi all

I restored a B&K 700 tube tester to working order by changing all out of spec capacitors and resistors, spraying the pots and calibrating it according to here:

https://tubesound.com/bk-707-tube-tester/

https://bama.edebris.com/manuals/b&k/700/

The 6BN8 is NOS (allegedly, DC voltage of 148 VDC checked out) and the #83 was replaced by a solid state replacement.

Everything checks out except that the DC voltage which should be 200 VDC is around 238 VDC in my case. I measured this with a Fluke 70 II multimeter.

When I test for example an EL84 in it, it drops to around 175 VDC.

Is everything ok or do I need to do further troubleshooting and adjustments? This is my first tube tester so I have no idea what I'm doing...

Thanks for any tips

[Roehrenonkel]

Hi agent_power,
 
what replacement for the 83 did you use?
How's the voltage with the original rectifier?
The 83 is mercury-gas-filled and has low regulation
over current (nearly constant voltage-drop of 15V).

Good luck

[agent_power]

Thanks for having a look

With the original #83 tube I get 232 VDC as measured on pin 3 and 8 of socket #29 when pressing the test button.

I replaced it according to the instructions of the PDF in post #1 with two diodes, two 10R resistors 1% and the fuse in between.

Interesting tidbit: I tested the original 6BN8 with it, and only one section died, other two were fine

[Roehrenonkel]

Hi again,
 
make sure you have set the line-voltage right.
Maybe you can get the voltage down with a (or two) power-zener(s).

I'm not familiar with B&K-testers,
have just a Tesla BM215A, a Kalibr L3/3 and a selfmade curve-tracer.

6BN8: two diodes, one triode - what died?
If you should need one more, i have one NOS and thousends of other US-tubes.

Good luck

[agent_power]

The voltage is regulated automatically by the tester, I'm running it on a variable transformer at 117 VAC until I know it's working as it should.. then I'll just use a regular step-down transformer. I'm also pondering of finding one that is small enough to hide inside the case so I don't need to lug one around.

Could it be that running it on 50Hz is causing problems? See:
https://antiqueradios.com/forums/viewtopic.php?p=1497612#p1497612
The tester is meant of 60Hz systems after all...

I could hook it up to a scope and have a look at the waveform.

6BN8: The triode for sure, upon further checking one of the diodes is also not working right
If I had known I could buy from the community I would have asked before Now I got some from the far east, they appear to work well

[Wallace Gasiewicz]

We run old tube testers (B&K among them) at the shop all the time, they work well on modern line voltage.  This is in US.
Putting in solid state diodes does up the voltage but that is what the 10 Ohm resistor is for,I do not think 50 or 60 Hz line freq makes a difference with the transformer in the unit.
The voltages on the tube is not exact in these testers, and is set with the switches. There are testers that do have adjustable plate voltage but they were much more expensive.   
B&K made good middle of the road tube testers, the kind a TV repair person would have.  They are not Lab Grade instruments. 
It should work fine with either  the diode rectifier or the tube rectifier. 
The voltage drop with load (the tube being tested) is to be expected, check the tube manual for the typical tube voltage of the tube being tested. 
You may wish to check the resistors in the various circuits, they may have changed value quite a bit.   
But I think it should work fine with the voltages that you have measured.

[agent_power]

Hello

Thanks for the sanity check. I have measured the resistors and only changed the ones out of spec (namely the two 82R and one 270R), some drifted but were barely within their tolerances. I'll go over them again next year.

Middle of the road tester is fine for me, I really just need it for repairs and restorations to weed out bad tubes and wanted to ensure it's not falsifying the results too much.

Some more things:
- I don't know if the line cord is original, but the earth lead was disconnected. Any recommended place where to put it, if necessary?

- If anyone knows what the latest tube chart is and where to download, please let me know. The one that came with the tester is from '72.

[Wallace Gasiewicz]

I do not think there is an Earth on the plug. I do not remember one and there is no Earth on the schematics.  Lots of equipment from this era did not have Earth and many homes did not have Earthed electrical sockets. 

For clarity the manual states 117-125 Volts and 50-60 cycle

[agent_power]

You're right, though my copy of the manual says 105-125 Volts, 50-60 cycle. The faceplate says 115 V 60 cycle.

Regarding the earth plug, someone must have changed the line cord during the life of the tester if it didn't come with a polarised one. If there is no benefit of earthing it then I'll just tuck the dangling lead away.

[floobydust]

I think the two diodes, two resistor replacement is not right, something I dislike. People add a 5W series zener to correct for the tube rectifier voltage drop. 10Ω is silly low value and does nothing much here for voltage drop.
Also, I don't see a filter capacitor- so the "200V" might be pulsing DC and multimeters do give different readings for average vs true RMS.
Nowadays mains voltage is high, the power transformer has less load with no #83 filaments to power, so I'd expect voltages to be high.
I remember back in the day the meter needle does move around with mains fluctuations.

[agent_power]

What does the Zener do differently than the two 10R resistors?

When looking at this post https://music-electronics-forum.com/forum/amplification/guitar-amps/maintenance-troubleshooting-repair/929305-b-k-precision-700-tube-tester-with-5z3-rectifier?p=929343#post929343 (I have configuration 4.) it doesn't seem to be bad practice doing it like this.

In any case, with the variable transformer set to 117 VAC, the original #83 in I still had higher than nominal voltage which suggest my measuring method must be the culprit. The schematics mention all but some voltages are measured 20kOhm/VDC and 5kOhm/VAC.

I don't have any such multimeters so I would have to find out how I could load down the plate voltage line with the correct amount and see if it changes anything.

[floobydust]

A mercury vapor rectifier has a voltage drop as well as internal resistance. Common practice for a guitar amp is different than test equipment.
Ultimately see how the calibration is - you don't want it reading say 15% high. I would check what a transconductance measurement gives compared to a bench measurement.
This tube tester is a bit awkward using pulsing DC not filtered DC.
#83 datasheet 1 is around 15V drop, curves around 123Ω from the slope, datasheet 2
Nothing even close to 10Ω.

[trobbins]

I'd agree with floobydust's comments, and similarly suggest your metering method accounts for the bulk of the discrepancy.  For example the rms of a nice sinewave is 0.707Vpk, but that assumes no flat-topping or other distortion, which is likely 11% higher than a vintage average meter - your measurement is 19% higher with ss equiv, and 16% higher with orig tube.  It's also worth noting that the tubesound webpage indicates plate voltage is 'a little above 200V' in the Mutual Conductance description, but the meter shows 195V in the test section with a Fluke 179.

[Wallace Gasiewicz]

Frequently the Manuals call for specific meters and mention ohms per volt or specify VTVM .  Here they call for a 20 Kohm per VDC, a modern multimeter is closer to a VTVM than the passive meters used in common measurements in this era of equipment. These meters would indicate a lower voltage than a modern multimeter.
They did mention a VTVM but in reference to certain AC measurements indicated by a dot.   
I am not quite clear about this. if I am mistaken. please elucidate.

[agent_power]

I'd agree with floobydust's comments, and similarly suggest your metering method accounts for the bulk of the discrepancy.  For example the rms of a nice sinewave is 0.707Vpk, but that assumes no flat-topping or other distortion, which is likely 11% higher than a vintage average meter - your measurement is 19% higher with ss equiv, and 16% higher with orig tube.  It's also worth noting that the tubesound webpage indicates plate voltage is 'a little above 200V' in the Mutual Conductance description, but the meter shows 195V in the test section with a Fluke 179.

I gather from this comment my Fluke isn't adequate enough anymore

With the SS #83 replacement still in and a scope attached, the results change drastically: 205VDC as measured by the scope. Also noted that my meter values have dropped aswell due to the scope's probe loading down the circuit (see attachments). Now we're at 'a little over 200 V'. With everything else checking out, would you agree this is withing specs again?

Frequently the Manuals call for specific meters and mention ohms per volt or specify VTVM .  Here they call for a 20 Kohm per VDC, a modern multimeter is closer to a VTVM than the passive meters used in common measurements in this era of equipment. These meters would indicate a lower voltage than a modern multimeter.
They did mention a VTVM but in reference to certain AC measurements indicated by a dot.   
I am not quite clear about this. if I am mistaken. please elucidate.

Mr. Carlson has done quite a good job explaining this: https://youtu.be/nOYWwIYH3bw?t=3444 - it's about having a load on the circuit to not blow out the multimeter needle in case you use a VTVM

[trobbins]

The reality is there is no 'spec' on that supply rail used for 'GM Plate test' - the schematic level of '200' is effectively an advisory level, especially as the tester has no fine adjust for mains voltage (it uses another technique to alleviate mains voltage variation).  So as floobydust recommends, it will be the Gm measurement capability, which for this tester is a bit wishy washy as far as absolute level - so will come down to consistency for a particular type of tube you may have tens or hundreds of, and how you interpret some that are replace versus good, or to grade valves in a batch as better than others.

You could measure the winding resistance of each of the 193V secondary windings, and then use PSUD2 simulation to assess the change in rectified voltage with loading, as a way of appreciating the change in reading with and without a scope probe connected.  To me, adding a 1Mohm scope 1:1 probe to a 10Mohm DMM load is not likely to see more than 1-2Vrms change, but you indicate the change with/without scope probe added was 'drastic'.

[floobydust]

On the schematic we have factory voltages taken by VOM, which is commonly RMS-reading, average-responding. Just like the Fluke 70_II (not true-RMS) reading for ACV and no idea about pulsating DCV compared to a panel meter.
"All voltages measured with 20kΩ/VDC 5kΩ/VAC" "* MEASURE WITH A.C. VTVM" {this is for "signal voltage cal R20"} - The Fluke DMM is much greater input impedance, something to note when working on old gear.
We have factory voltages: Transformer "193-0-193 VAC" yet "200VDC" so people can chew on that.
I do know that mercury vapour rectifiers have absolutely terrible reverse-recovery times, in the blue glow which might explain a bit more. Odd that OP's scope trace shows clipping.

But I did find B&K Model 707 factory changed from #83 to two 1A 600-PIV 1N4006 rectifier diodes. They added 56Ω resistor at the panel meter, I think the panel meter is 1mA 100Ω movement. This might be for the higher voltage? Also from 1MEG to 2MEG pot for SHORT SENSITIVITY.

Many old test equipment schematics here: https://www.pacifictv.ca/wanted.htm and 707 tube chart too. It looks like near the same as the 700.

[agent_power]

But I did find B&K Model 707 factory changed from #83 to two 1A 600-PIV 1N4006 rectifier diodes. They added 56Ω resistor at the panel meter, I think the panel meter is 1mA 100Ω movement. This might be for the higher voltage? Also from 1MEG to 2MEG pot for SHORT SENSITIVITY.

Many old test equipment schematics here: https://www.pacifictv.ca/wanted.htm and 707 tube chart too. It looks like near the same as the 700.

I think the different resistors come from different transformers or meters, there's a warning on tubesound.com about this.

I'll probably not use 707 tube charts with the 700 because I couldn't find the exact differences and I don't want to blow up a tube because the sockets are wired differently....

The latest tube chart specifically for the 700 is here: https://bama.edebris.com/manuals/b&k/700/

Anyway: With the original #83 in and taking the same measurement as before: 200 VDC with a *cold* #83, around 180 VDC after warming the tube up for 30 min - where also the clipping seems to have gotten less worse.

The reality is there is no 'spec' on that supply rail used for 'GM Plate test' - the schematic level of '200' is effectively an advisory level, especially as the tester has no fine adjust for mains voltage (it uses another technique to alleviate mains voltage variation).  So as floobydust recommends, it will be the Gm measurement capability, which for this tester is a bit wishy washy as far as absolute level - so will come down to consistency for a particular type of tube you may have tens or hundreds of, and how you interpret some that are replace versus good, or to grade valves in a batch as better than others.

You could measure the winding resistance of each of the 193V secondary windings, and then use PSUD2 simulation to assess the change in rectified voltage with loading, as a way of appreciating the change in reading with and without a scope probe connected.  To me, adding a 1Mohm scope 1:1 probe to a 10Mohm DMM load is not likely to see more than 1-2Vrms change, but you indicate the change with/without scope probe added was 'drastic'.

From which points do I take these measurements? As you said, the meter is wishy washy and it does what I want it to do, weed out bad tube from good ones. I'll probably learn with experience what the results mean from using it and as I don't need to grade them at precise Gm levels I think we can conclude that the circuit is working as intended.

I have never used PSUD but I attached a file in case anyone wants to play with it. The resistances were as follows:
RED - 327R - GND - 226R - BRN / RED - 625R - BRN
With circuit transfer switch on: RED - 195R - GND - 275R - BRN / RED - 471 - BRN

I'll probably still use the #83 SS replacement, it draws 30W less, doesn't bake the components inside and allows me to hide a smaller transformer inside the case for 230V operation

[floobydust]

Looked at it some more and you can see how BK's design evolved from the 700, 707, to all solid-state 747. I thought the same power transformer is used among them.
I don't know if they compensated for sag with the #83 rectifier. How much plate current will the tester push, say a 6L6 is it up to 200mA say?
I get #83 internal resistance around 120Ω based on the slope of datasheet graphs. So I still don't like the 10Ω there in the replacement but BK had zero in the 747.