Featured
Latest
Testing Valves
 
Share:
Notifications
Clear all

Forum 141

Testing Valves

33 Posts
5 Users
0 Likes
2,652 Views
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Hi Folks,

I've just completed building a simple valve tester to test a range of audio valves (12ax7, EL34, EL84, 6V6, 6L6 etc.) and I would appreciate some advice when deciding which valves are good or not, please.
The tester puts 250v an the anode and a little less(via a 1K ) on the screen. The cathode goes to earth with an ammeter in the anode supply and I can vary the grid from 0v to -28 volts
I made a chart from the various valve data on the net to use as a guide but the results do seem to vary quite a lots from the published figures - which seem to be quite definitive with no tolerance.
Testing the 4 new El84's I have with -7.5v on the grid I get anode currents ranging from 39mA to 33mA from so-called balanced valves.

Want sort of anode current range should I be looking for when deciding if a valve is on its way out?
Regards
Geoff.

 
Posted : 28/05/2013 1:09 pm
Valvebloke
(@valvebloke)
Posts: 894
Honorable Member Registered
 

This is not a question with an easy answer. A valve which doesn't work in one circuit may well be fine for another. And whether it's 'on its way out' or not depends not only on how it measures now but also on how it measured when it was new. If its emission now is only half what it was when it was new then the fact is it's fading, and maybe fast. But if it started out with very high emission then it may still look OK in a one-off test.

There's some very good advice about valve testing in general on Jac van de Walle's site http://www.jacmusic.com/ . When you get to the site:

1. Choose the US/English option in the left hand menu
2. On the next screen scroll down the left hand menu and click on the Tech Corner section at the bottom
3. On the next screen scroll down the left hand menu and click on the Tube Tester Collection (section 7)
4. On the next screen scroll down the page a little way to Contents and click on section 3 there - How to test tubes, on various testers.

VB

AmpRegen http://www.ampregen.com

 
Posted : 28/05/2013 4:28 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Hi Valvebloke,
Many thanks for directing me to this very interesting site. Having read through a good deal of the site information, I now understand why I use to just replaced a suspect valve with a new one from stores and moved on to the next piece of faulty equipment awaiting repair!

I'm also beginning to question why I built a valve tester in the first place - just for the hell of it I guess!
One question that building a tester has posed is; what does "balanced" mean? Does this mean the anode currents are within a specified range or is the Gm within a particular range. Suppliers of matched valves don't seem to specify any figures !!! Does it really matter as I've seen some guitar valve amplifiers with a difference of 25% between the anode currents of a push pull output stage with no sign of cross-over or other distortion on a scope. I can always remember my old tutor a college ( good memory ha!) saying that if you can't see distortion on a scope you can't hear it.

Regards
Geoff.

 
Posted : 28/05/2013 9:20 pm
Valvebloke
(@valvebloke)
Posts: 894
Honorable Member Registered
 

Most sellers who claim to sell 'balanced' or 'matched' valves just match on anode current at one particular operating point (usually the one in the valve tester manual). This is better than nothing but, frankly, not a great deal better. Some sellers, but not many, can be prevailed upon to test the valves at the operating point you want to run at. So, for example, they will match KT66s for a Quad II at 330V anode-cathode and -28V on the grid, not 250V a-k and -15V on g1 which is the test point GEC give in their data sheet. A very few will match on gm as well as Ia. But gm tends to be less important since in the sort of amp where you care about distortion there's usually enough negative feedback to balance up two otherwise decent valves which happen to have unmatched gm's. Things are a little different in guitar amps where you quite often find bias controls allowing you to set Ia but where there's usually no global NFB. Of course distortion is not really a problem in guitar amps. It's an opportunity :=D !

VB

AmpRegen http://www.ampregen.com

 
Posted : 29/05/2013 12:40 am
Refugee
(@refugee)
Posts: 4055
Noble Member Registered
 

The guitar guys want the clipping to set in evenly and the sellers will match the valves to do this and will not bother about the low volume settings.
At normal domestic volume levels it matters much less and most listening will be at lower volume.
The shape of the getter does not make a difference I can detect on the set with amps like the one in my avatar however it is a painful experience listening to them close to clipping.

 
Posted : 29/05/2013 2:41 am
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Hi Guys,
Many thanks for the great advice.
Best regards
Geoff.

 
Posted : 29/05/2013 9:58 am
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Hello Geoff
I am embarking on a similar project as yours. Mainly to learn more about the characteristics of valves and to test the growing number of audio valves which seem to be accumulating by my son replacing the valves in his guitar amps.
Is there any chance you can post your circuit.
Thanks
Mike

 
Posted : 30/05/2013 2:43 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Hi,
Its only hand drawn at the moment but will try and re-draw soonest.

Regards
Geoff

 
Posted : 30/05/2013 10:17 pm
valvekits
(@valvekits)
Posts: 780
Honorable Member Registered
 

I would imagine it might resemble one of the first Lab tutorials we did on valves.
We used bench power supplies to plot a family of grid curves for the ECC83 which I think is probably the best way to get acquainted with valves.
I wonder how many of us still have those college notes locked away in the attic.
Geoff's set-up would probably be more useful if the HT power supply could be made variable by say (ahem) a simple power Mosfet circuit.

Eddie

 
Posted : 31/05/2013 11:14 am
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

I'm building this. Needs no expensive hard to get transformer. Uses a 19V to 20V Laptop PSU brick
http://www.dos4ever.com/uTracer3/uTracer3.html

 
Posted : 31/05/2013 11:25 am
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

hi Michael
I was looking at the Utracer today, are you sourcing all the parts yourself or buying the kit?

I have opted, for the moment, to try building something like the "Sussex" valve tester. I have just finished drawing a layout for the "Sussex" leakage and grid circuits on a piece of veroboard, it took me literally all day yesterday.
I plan to start off with a simplified HT circuit and will probably modify it to the "Sussex" style anode and screen circuits later.

 

 
Posted : 31/05/2013 11:45 am
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Hi Folks,

Although I've been able to re-draw the circuit of my "simple" valve tester using a software package, I've taken a photo so if anyone would like a copy, please email me as the file is too big to attach. Having used the tester on my small stock of valves, the stability of the HT supply is not as good as I'd hoped. The mains transformer was purchased 2nd hand and I don't think its fully up to the job. However I'm not changing it now because when you look at valve data curves, anode current does not seem to be that effected by a small fall in HT when the valve is fully driven.

For reference, here are the results from (believed good) EL34 I tested today:
Valve Bias Anode Current
1 -15v 48.1mA Electro Harmonic
1 -0v 118mA "
2 -15 49mA "
2 -0v 114mA "
3 -15 48.6mA NOS Mullard
3 -0v 122mA "
4 -15v 46.4mA NOS Mullard
4 -0v 119mA "

I would appreciate any comments on the above, please.
Regards
Geoff

 
Posted : 31/05/2013 1:15 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

A donor who wishes to remain anonymous gave me the kit as a present.

I am going to add some customisations.

The cost of getting valve sockets is significant. Make your tester in a big enough box to add sockets. (use ferrite bead near the valve bases).

You need as a minimum these four separate supplies
1) +HT1 12.6V to 250V Anode Supply.
2) +HT2 12.6V to 250V Screen Grid Supply. You must test at makers suggested g2 volts or Ia vs Vgk will be madly out.

These two HT supplies can be steps such as 12.6, 22.5, 45, 67, 90, 120, 200, 250 etc. If the current metering is calibrated (See AVO VCM163 explanation) you can use unsmoothed halfwave supply and adjust mains input selector for your local supply. As soon as you try to add smoothing you need a regulator. A suitable "robust" solution is EL34 or 6146 as a cathode follower with DC feedback and pot to set HT voltage. Two such circuits are needed.

3) -Control You need 0V to -30V for grid. Or even +5V to -40V if testing some valves for +9V to +14V HT or Power valves intended for Class B / Class C Bias. You can use unsmoothed halfwave supply if the metering is calibrated. The VCM163 modulates the grid with a fixed higher frequency signal (set at lower preset level for high gain valves) and uses a filter with peak detector on the Anode to measure actual signal gain as well as display of Screen Grid and Anode static current.

4) Filament Heater
0.5V to 120V for Heater at up to 6A at 4V. Only mA at 0.5V to 1.6V and above 20V.
Here you could do variable 0.5V to 1.6V and 1V to 3.2V for battery valves up to 100mA, 2V, 4V, 5V, 6.3V & 12.6V at several amps for older valves or parallel mains valves and adjustable current regulator 23mA to 350mA for Series valves.

When testing valves you need stable known recommended filament/heater volts or current and also you need to be able to exceed normal voltage / current to rejuvenate valves. Also you may need to characterise the valve at the extreme of supply voltages in target equipment (e.g. 230V to 250V mains variation or 1V to 1.6V battery variation).

A set of sockets and patch leads to connect the heater/filament, 0V, HT1, HT2 and -Control voltages. Heptodes, Triode Hexodes, Octodes etc will need multiple HT2 and OV.

g3 isn't always at zero volts. A deluxe touch is +/- 12V variable range on g3.

There are four main usages:
1) Pass / Fail testing to decide if apparatus or valve is faulty (do you need to order a valve?)
2) Rejuvination. Usually you set g1 just to cut off and increase heater voltage. As emission rises you keep turning down g1. g2 typically at 1/2 anode voltage.
3) Matching valves for Push pull or parallel use. Need to set g2, g3, anode and g1 to target apparatus quiescent conditions.
4) Generating characteristic curves to either understand a valve (Tetrode in a Negative Resistance Oscillator) or have better data for a design or decide if a replacement will work. For example some valves LOOK as if they are different gm and Ia vs Vg quiescent conditions, but in fact the maker has used quite different Vg2 (screen) and Anode. E.g. ECH81 and ECH83 are the SAME valve, but one has curves for 12.6V g2 and Anode and the other for 150V to 200V! You can see the massive effect of leaky decoupling cap killing gain on a Pentode or Tetrode g2 (say 10V instead of 150V) and effect of too much HT (e.g. 90V instead of 45V to 67V on a DK40 Osc /screen grid supply.

Another useful feature is to connect a capacitor leakage tester (say 200V to 350V) or circuit of a flash gun via a uA meter (use 2 x 1M series resistor for Electronic flash supply) with protection diodes across it to test for inter electrode leakage with heater off and on.

 
Posted : 31/05/2013 1:19 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Hi Geoff
Thanks for your circuit, it looks functional, yet simple and easy to build and would make a good basis to add to as and when.
I especially like the control panel and cabinet, is that cabinet from your best cuttlery set? :D

If you would like me to reduce the images and post on here, let me know.

Thanks
Mike

 
Posted : 02/06/2013 8:19 am
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Hi,

If you could post the images I would appreciate it.

In terms of costly parts, I purchased most from Ebay, in particular the valve bases were under £3 each from Hong Kong and of good quality. If fact as I was not in a rush, most parts were purchased from Hong Kong and the total bill was under £40 - excluding the vintage box which I picked up from a car boot for a fiver. I do like the wooden box finish, it give the project that retro look !

As you can see from the circuit I don't have any internal valve short circuit / leakage detection. Therefore any suggestions would be appreciated,please.
Regards
Geoff.

 
Posted : 02/06/2013 12:57 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Here are the photos that Geoff sent me.
Mike

 
Posted : 02/06/2013 11:17 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

All he needs is insides of old Film camera flash or new £5 Disposable Flash Kodak film camera
1) Remove Flash tube
2) Remove 300V etc Flash Electrolytic and replace with 1uF 500V Met poly
3) Put 2 x 1M in series with + of cap to electrode under test and - of cap to meter
4) Other end of meter to other electrode. Put a diodes across meter to protect

Ideally a 20uA or 50uA meter. This will also work for the mA Anode current with a shunt.

 
Posted : 03/06/2013 12:16 am
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Is the idea of the diodes across the meter to "shunt" any surge voltage which is greater than the forward voltage drop of the diodes.
Thanks
Mike

 
Posted : 03/06/2013 10:14 am
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Yes. it's on the AVO VCM163 and many testers.

 
Posted : 03/06/2013 10:57 am
valvekits
(@valvekits)
Posts: 780
Honorable Member Registered
 

Hi Geoff
Your tester is pretty much how we did the lab tutorials except we could vary the supplies and had a voltmeter connected to the anode.

A lab experiment from a long time ago…..

Experiment to plot the characteristics of a pentode valve EF37 or equivalent
Method
Connect the suppressor grid (G3) to the cathode.
Adjust the screen grid (G2) voltage to 100V
With the control grid (G1) at zero, measure the anode current as the anode voltage is adjusted between 10V and 100V in 10V steps.
Repeat for different values of Vg1 -0.5, -1, -1.5, -2
Plot a set of mutual characteristics.

With Vg1 = 0, make Vg2=Vg3=60V and measure the anode current as Va is varied between 0 and 120V.
Plot the anode characteristics for these values of Vg2 & Vg3 and note it's resemblance to that of the tetrode.

If you forget the bit about connecting G3 to the cathode, that might cause some strange results.

Eddie

 
Posted : 04/06/2013 10:21 pm
Page 1 / 2
Share: