The "Sussex" Homebrew Valve Tester.

[jimmc101]

Hi Bero
For the most sensitive range:

Input = 200mVrms = 180mV average (modulus)
First stage gain = (5k6/2k7)+1 = 3.07
Second stage gain = 2.35 (see previous post)
So input to third stage = 180mV x 3.07 x 2.35 = 1.3v average

As you say current is 1.33mA

Thus value of (330R + pot) required is 1.3v/1.33mA = 977R
So pot will be set to about 977-330 = 647R or roughly one third of 2K.

It all sounds about right.

Jim

[Bero]

Thanks Jim,
I was wrong because didn't calculate with first and second amp.

[Top Cap]

Well I am still awaiting more replacement parts due to a mild explosion owing to the fact that the current limiters were not operating correctly. The 10 Ohm resistor used for the Gm meter was blown open and consequently it took out the Gm meter. The MPSA42 on the Anode Volts circuit is open circuit as are the back to back 15V protection Zener diodes. The MPSA92 pre-regulator was short circuit which probably accounted for the high HT I had, lesson learned here as the spice simulation suggests there should be just about 300V and so I shall be checking for that. I did find two errors, the MPSA92 emitter resistor was 47 Ohms and not 470 Ohms. Probably trying to read resistor colours through old eyes complete with cataracts. The other problem was that the Heater/Cathode leakage LED lit at the same time as the Heater Continuity one. Once the function switch was turned away from Heater Continuity both LED's extinguished. I found that I had wired the Anode of the Heater/Cathode leakage LED to the heater rail feeding the pin selector switches when it should be connected to the wiper of the heater voltage selector switch. Thus it is isolated when RL2 operates whilst checking continuity. So if anyone else has this problem then this may be the answer. I would like to reduce the value of C13, the 22uF wired from the MPSA92 collector and hope I can get it down to 2.2 or even 1uF.
Was there any particular reason for it being connected here? The spice simulation shows charge time to be very long but of course the simulation may not be right, it has fooled me before.
Also wondering if a diode across the MPSA92 collector/emitter might reduce the chances of damage from reverse voltage on switch off, though once again simulation shows this as not being a problem so may not be necessary?
The total failure of the MPSA92 is a bit of a worry and I may have had one of those devices with reversed Collector and Emitter connections as shown in the attachment. Will definitely put them on the hfe tester when they arrive!
The new IRF830 FET's are in but like me, you will have to be a bit patient while the other spare parts are in transit.
I must reiterate though the joy of having a plug-in assembly as I would have found the removal of the pcb's a wretched chore otherwise.
Les

[Andymic]

Glad to see this thread back 'on' topic.

Les,
Is your layout diagram correct re the MPSA92?

Andy

[Top Cap]

If you are using Nat Semi's or Fairchild's otherwise you will have to check.
Of course there may be an error on the data sheet of the other manufacturer.
Probably a red herring but I think worth a check anyway.
Les

[Bero]

New chips fitted, thing sprang to live again.
This reading is actually 63,9mV.
My LCD isn't so "sophisticated" like Mike used, I've must do something to give at least decimal point.
This is the first trouble to solve.

Second is a reason why chips burning.
If I want to measure AC volts across 100R resistor in anode circuit,
I've thinking that would be sufficient to put DC blocking cap to only one side of the meter input.
Just to break circuit to prevent DC voltage passing through the meter (Figure 1).
But chips burned in this case, I must put blocking cap to both sides to prevent burning (figure2).
I know there is usually 250V both side of the 100R resistor,
but anode circuit isn't ground sharing with meter supply.
Why this happen?
Can those caps produce influence to meter precision?
Anny suggest for caps value?

[Top Cap]

Well some good news and a query! I fitted the IRF830 FET's and replaced the other components as required but this time, as I should have done before, I wound the input up on the Variac. The pre-regulator settled at 297V and I was happy with this. Checking the Anode and G2 supplies gave me fully controllable voltages on both down to 0V. I repeated the exercise with a very low load on the supplies and they did indeed both current limit so I am a happy man now. The query? Well I came to test an EZ41 but when I inserted it into the socket it flashed violently inside and the Tester gave a horrible buzzing sound due to the minus supply being shorted. This happened with the valve selectors properly set up and with the Tester set to Heater Continuity. The valve had a shorted cathode/heater and thus was shorting out the minus supply which is why I was treated to fireworks. As a quick fix I replaced the 10k heater continuity LED series resistor with a 4k7 and placed another 4k7 in the 'hot' line by bridging a track cut made between the minus supply and Relay 2b NO contact. The only problem with doing this is that a shorted cathode/heater makes the Heater Continuity LED extinguish. I suppose the only real answer would be to add another relay such that the cathode line is open circuit whilst the heater continuity is being checked. Alternatively, set all pin switches to 0 before doing continuity tests and set them when the function switch has been moved to Leak positions. Adding an extra relay however will prevent an accidental switch to Heater Continuity causing a short circuit on the minus supply and more importantly, any damage to a battery valve whose heater MUST be isolated from the cathode line in the Heater continuity switch position. It may be possible to fit the Heater continuity LED and 10k resistor in the 'hot' line as this will also provide a limiting of current and thus protect not just the minus supply but also any battery valve filaments, as per attached pic.
Has anybody else had any problems with the BUZ80's or perhaps my pair were faulty?
Les

[Top Cap]

Further to the above, I think the attached may work. I know that my design is slightly different to other folks but you should be able to see how I have made R28 now perform the role of current limiting for both the Heater Continuity and Heater/Cathode LED's resulting in the removal of R29 and some slight re-wiring around RL2. If you have added a 50 way connector then the scheme allows the modification without having to add or disturb the wiring to this connector. In order to do this the circuit path has been reversed which should be of no consequence to its Heater Continuity operation. Just having another coffee before getting stuck in with the mod. Go or Blow, I will let you know .
Les

[Top Cap]

Nope! A heater cathode short now keeps RL2 self latched, back to the drawing board!

[Top Cap]

Simple answer, take RL2b pin 14 to its own separate 0V, I used main board pin AC.
Unit works first class now, no nasty sparks if the valve has a direct cathode to heater short and yet the cathode/heater LED picks up the fault when you select leak tests etc.

[Andymic]

Les
Did you update the layout at http://homepage.ntlworld.com/lez/Sus...nstruction.JPG ?

Andy

[Top Cap]

Hi Andy, not yet but will do. My VT passed another, if not final, milestone by successfully testing a DF33 battery valve with a 1.4V heater. Well pleased.
Les

[Top Cap]

Updated the sketch, just hope I have made all the changes correctly but it looks right. Still confused over those BUZ80's I had and very interested if anyone has similar problems. I have been assured by colleagues at work that as these devices are meant for high voltage use then the Zener diodes should not be a problem so it is all a bit of a puzzle.
Les

[Guitarist28]

Hi All,
I thought that I would give a progress report on my case (see attached) which I know think might be a little on the large size compared to the others I have seen. Anyway, so far its constructed out of 3mm aluminium sheet and I would say is approximately 2/3rds avo vcm sized or there abouts.

I don't think that I am going to have to much trouble with space for the internal wiring!
In hindsight I should have bought a ready built enclosure as I could have finished this project by now.

Rob

[msmpegr8]

Hi Gents,
Newbie to the forum - Looks like a very nice tube tester, Mike!
I'm slowly working on a S. Bench TT. Once I study Mike's, I'm sure it'll give me some ideas...

Good luck on your projects,

Mike 8>)
If there's no sound in a vacuum, where'd the music come from?

[Bero]

Mike (swordholder),
Can you please comment basic difference between your design and S.Bench RAT tester?
Like I see, S.Bench design has current limiter,
variable from about 100 microamps to 50 mA via a potentiometer controlling a constant current circuit.
There is a not possibility to set negative grid voltage.
Your design has potentiometer allowing adjusting grid voltage.
Can you give some comments please?
Regards

[swordholder]

Hi Bero,
Not too sure about all the aspects of SB valve tester. I based mine roughly on the AVO VCM163 with which I am familiar, hence the variable grid voltage and the audio signal superimposed on it to measure GM.
Mine differs from the VCM163 by using DC on the anode and screen whereas the AVO uses AC, which involves a multi tapped (and expensive) mains transformer.
The advantage of using the VCM is that books are available giving full set information for many valves.
Hope this answers your question.
Regards
Mike

[Bero]

Thanks Mike,
Maybe answer on my question is here, I found those definitions:

Transconductance (gm) is defined as the incremental change in plate current for an incremental change in grid voltage, with all other parameters (plate voltage for example) held constant.

This is doing by your design, you achieve valve biasing by setting grid voltage.

Except (gm) measuring, SB design allows (mu) to be measured.
According to definition:
MU is defined as the incremental change in plate voltage for an incremental change in grid voltage, with all other parameters (plate current) held constant.
This is the reason why SB design has current limiter, addervise couldn’t be (mu) measuring achieved.

I didn't noticed this before, until study a little your project, I'm a little confused now.
All seems to look OK when I've tested triodes like ECC82, 83, 85 for gm/mu.

But when I've try to test EL84, and set initial parameters (plate on 250V, g2 on 250V, current 45 mA)
grid voltage became a lot in positive respected to cathode.
By current less then 1mA, g1 is almost 25V negative to cathode.
After increasing current to approx. 5mA, grid to cathode is 0 volt.
All current increasing after this point goes to grid became 20 and more volts positive respect to cathode.
Can this manner of valve biasing destroy valve under the test?
Your design allows setting negative grid voltage according to book specif. and that's ok,
but I'm confused now, how SB design doing this?
Any comment welcome.
Regards

[Top Cap]

I had problems testing EL84's and EF183's and it was due to parasitic oscillation which you can check by using a scope on the control grid. I tamed mine with a 1nF disc at each end of the valve pin selector wiring (between G1 and Cathode) though I am tempted to wire this in screened lead, just for the Grid 1 position (switch setting - 4).
My other problem is that after testing a valve and switching back to Heater Continuity. the leakage LED lights at the Function switch position for Anode/Screen to 0V leakage. It slowly fades out as the valve cools and sure I know it is due to conductance within the valve, I am wondering if I could somehow reverse the test voltages so that the -45V appears on the Anode?
At least knowing what it is I am not too concerned when it lights after testing.
Les

[G8UWM-MildMartin]

If the grid G1 really does go positive when fed from a negative voltage, I agree with previous comments that it might be oscillating or, more likely, exhibiting a common failure mode of physically small output valves.
It would be worth trying a good valve to work out which it is!
Prolonged testing under these conditions should not really cause any further damage to the valve if the anode and screen dissipation are kept within ratings.
I suppose there's not much to lose by grossly over-running it for a few tens of minutes to see if the getter will do anything.