Hi Retailer!retailer said:
Mmm... It sounds interesting. I´ll try modifying R13 and I´ll make some test with CR17 with different voltages. I´ll let you know.
About the meter...poor results. I´ve in AC200mV sine wave on TL072 gate3 and get 750 mV at gate 1 and 2Volt sine wave at pin 7, so lokks like correct but on gate 6 /L071 I got only about 50mV (oscilloscope screen) but fully distorted signal like almost a square one and not output so, I made some mistakes. I´ll check it this eek end.
On the oder hand about tru-rms voltmeter, AD636 is a very hard to find component. I´ve got AD536 TO-100 and make some test on the schematics. Good luky in moding your son's recording studio and have a nice weekend!
jmsegui@ya.com
Re: RAT Tube Tester - All Circuits/Schematics
Hi Dominic, I´m posting schematics of my RAT version. I´ve made some mods and use banana pig-tails like SOFIA tester. Regards
jmsegui@ya.com
DomKo said:
Hi Dominic, I´m posting schematics of my RAT version. I´ve made some mods and use banana pig-tails like SOFIA tester. Regards
jmsegui@ya.com
jmsegui said:
Here's how the AD536 stacks up to the LT1966 -- fwiw the LT1966 is about 1/2 the price and has a much, much lower error term.
An externally hosted image should be here but it was not working when we last tested it.
the residual errors in the LT1966 are close to zero.
I have AD536, 636 and 737
jackinnj said:
Thanks a lot for your info... you are absolutely right but I´ve to managed the components I have on hand. Nevertheless we are talking about a tube tester rather than a precise lab equipment (sorry) so I´m wondering whether using TRUE rms voltmeter is somthig like hunting birds by a12´cannon. In my oppinon a precision AC voltmeer would be good enougth but a DIY True-rms voltmeter is a deal. Cheers
jmsegui@ya.com
jmsegui said:
Your point is well taken, you can't keep everything on hand.
Even a good averaging AC meter requires a dual rail-rail opamp and an averaging filter -- and if you go this route you really need an opamp which is going to be acurate near the rails. Here's an idea which appeared in EDN:
An externally hosted image should be here but it was not working when we last tested it.
No diodes is a big plus, if you don't have a rail-rail opamp you need dual supplies.
I was looking around the internet for a pc tube tester came across the RAT thought it was great, but would be nice if it was attached to a computer. Then you can the computer do the switching and formulate a curve. Then i ran across this site about a pc tube tester that actually shows a curve as well as other stuff. just thought it would be interesting to share.
http://www.tubesontheweb.com/matic.htm
http://www.tubesontheweb.com/matic.htm
I have a Mercury 1000 that I had problems with when I got it. I ended up building a small test set to measure Gm of the 6BA6 so I could calibrate the Mercury 1000.
I ended up deciding to build a more general purpose test set for parametric measurement of tubes. I plan on using adjustable current source in the cathode circuit, adjustable plate and optional screen supplies. I am looking at the AD629 with 1ohm sense resistors for plate ac current measurements, and the LT1966 (just got two samples) for RMS conversion to get Gm from the plate current.
It looks like several of you are further along than I am (I'm still collecting parts).
The RAT tube tester looks interesting.
I ended up deciding to build a more general purpose test set for parametric measurement of tubes. I plan on using adjustable current source in the cathode circuit, adjustable plate and optional screen supplies. I am looking at the AD629 with 1ohm sense resistors for plate ac current measurements, and the LT1966 (just got two samples) for RMS conversion to get Gm from the plate current.
It looks like several of you are further along than I am (I'm still collecting parts).
The RAT tube tester looks interesting.
I just finished calibrating 2 Mercury 1000. (I repaired the first one and helped a friend repair the second one himself). They also come with a 6AT6 (ebc90) and it is more stable / easier to calibrate. The calibrate instructions for the 1000 cannot be found on the web but the 2000 can. Just substitute socket 12 for socket 8 mentioned in the test procedures.
The RAT is very good but I feel there is a logic error in the design, I have just started collecting parts for it. It is a class IV design (laboratory class). Have a read of the chapter on tube testers in the book Tomer_1960_Getting_the_Most_Out_of_Vacuum_Tubes.pdf which can be found on the web.
My problem with the RAT tester is that the g1 is at ground potential which means that the cathode finsihes up far above it (when testing power tubes it can go up to 48 V for the 6146). What needs to be changed is that the grid goes negative and that the cathode stays at ground potential. From the forementioned book I got the idea to remove the 100 ohm resistor across which the measurements are taken and replace it with a choke. Then there will be no DC potential / voltage dropping happening and the plate voltage is the true plate voltage which makes for better correlation with the test reuslts in tube handbooks.
Still plenty of these kind of mosfets around in this mart of the world, most expensive items are the two transformers.
I also disagree with the use of true RMS meters: since the input is a pure sine that is an unnecessary expense. An ordinary AC meter is sufficient and we need to use an amp meter.
AmadeusMozart
Why use RMS converters?
In my case I have plenty of digital panel meters that only display DC.
So, if I convert the ac to RMS, I can drive the digital panel meters.
For current readings, I plan to use a 1 ohm sense resistor and scale the voltage with op-amps. Again I can drive the digital panel meters.
In my case I have plenty of digital panel meters that only display DC.
So, if I convert the ac to RMS, I can drive the digital panel meters.
For current readings, I plan to use a 1 ohm sense resistor and scale the voltage with op-amps. Again I can drive the digital panel meters.
Yes, that will work just as fine. In my case I am still sourcing some stuff, hardest to get are the mpa transistors in a low quantity. Checked out Mouser but their shipping charges to NZ are outragious. Minimum number to get locally is 50...with a $10 handling fee plus shipping costs.
Have now got all the important parts and have started to study the schematics indepth. Am now understanding how the negative supply for the CCS works and am uncomfortable with the configuration in a dynamic surrounding like in a tube tester. I can understand why Steve Bench took this approach but I am going to attempt to improve on an already good design.
There are a couple of reasons for me being uncomfortable and if others want to chime in with their opinion I'll appreciate it. After all there may be errors in my logic.
First reason is that if the sink voltage is negative then you'll finish up with a positive voltage on the grid. A large number of tubes will start drawing current in that situation and you get on a slippery slope of where the grid will be destroyed with grid current. The start up may well mean some capacitors need to be brought up to the current voltage level and that period may be dangerous. This can potentially be overcome by having a simple diode from ground connected to the CCS so when it crosses it will not go more than 0.6V positive. (easily enough done)
Second reason is that if you are matching tubes then you want to match both the static DC current and the amplification factor. This means you want to be able to set the grid voltage. This is not possible and is more serious.
Third reason is that with power tubes you have the need for a large bias and have at the same time a high plate voltage. Then current sink has the characteristic that when the bias gets larger then this voltage is taken away from the maximum plate voltage so you finish up with a lot less than you'll need. I find this an even more serious drawback, don't want to loose 40 volts or more just for the CCS (plus more heat generated internally in the tubetester).
The fourth reason is that you want to be able to use manufacturer's tube specifications and the RAT tester measures the plate voltage against ground. However ground is not the same as the cathode potential ( what is used by the manaufacturers specs) and hence you are measuring against a varying level with varying results.
I'll be shortly prototyping this on a breadboard but will change a few setups: I'll make the grid voltage adjustable and have the cathode at ground level. Likely I'll build an adjustable current limiter into the plate supply and similarly into the screen supply. In due time I will report on this but activities are likely to take place over a few months due to other things I have on my plate.
There are a couple of reasons for me being uncomfortable and if others want to chime in with their opinion I'll appreciate it. After all there may be errors in my logic.
First reason is that if the sink voltage is negative then you'll finish up with a positive voltage on the grid. A large number of tubes will start drawing current in that situation and you get on a slippery slope of where the grid will be destroyed with grid current. The start up may well mean some capacitors need to be brought up to the current voltage level and that period may be dangerous. This can potentially be overcome by having a simple diode from ground connected to the CCS so when it crosses it will not go more than 0.6V positive. (easily enough done)
Second reason is that if you are matching tubes then you want to match both the static DC current and the amplification factor. This means you want to be able to set the grid voltage. This is not possible and is more serious.
Third reason is that with power tubes you have the need for a large bias and have at the same time a high plate voltage. Then current sink has the characteristic that when the bias gets larger then this voltage is taken away from the maximum plate voltage so you finish up with a lot less than you'll need. I find this an even more serious drawback, don't want to loose 40 volts or more just for the CCS (plus more heat generated internally in the tubetester).
The fourth reason is that you want to be able to use manufacturer's tube specifications and the RAT tester measures the plate voltage against ground. However ground is not the same as the cathode potential ( what is used by the manaufacturers specs) and hence you are measuring against a varying level with varying results.
I'll be shortly prototyping this on a breadboard but will change a few setups: I'll make the grid voltage adjustable and have the cathode at ground level. Likely I'll build an adjustable current limiter into the plate supply and similarly into the screen supply. In due time I will report on this but activities are likely to take place over a few months due to other things I have on my plate.
Last edited:
I'm still working on my tube test set design.
I'm considering the following basic requirements:
1. Test 7 pin, 9 pin, and octal (optional) tubes of the configuration Triode, Tetrode, Pentode, and combination tubes (Dual Triode, Triode/Pentode combination, etc).
2. Plate voltage Supply variable from 100V to 350V, for Gm Measurements. The voltage supply is current limited to 100mA for tube protection. This should be adjustable to help prevent damage to smaller tubes.
3. Plate ‘Current Source’ Supply variable from 0.1 to 100mA for Mu measurements.
4. Screen Voltage Supply adjustable from 100V to 350V current limited to 10mA for tube protection.
5. Grid Voltage Supply adjustable from 0V to –30V with current limit to 1mA for tube protection. No positive Grid drive? Are there tubes that require positive screen drive during testing?
6. Cathode directly connected to ground.
7. Voltage to Plate, Screen and Grid measured WRT ground and available for displayed.
8. DC Plate Current sensed across a 10 Ohm 1% (or better) sense resistor.
9. AC Plate Current sensed across a 10Ohm 1% (or better) sense resistor for determination of Gm.
10. AC Plate Voltage sensed directly at the plate with the Current Source supply driving it for determination of Mu.
11. DC Screen current measured across a 10Ohm (1% or better) sense resistor.
12. 100mV RMS 1KHz sine wave source for driving the grid to facilitate Gm and Mu measurements.
Comments please.
I'm considering the following basic requirements:
1. Test 7 pin, 9 pin, and octal (optional) tubes of the configuration Triode, Tetrode, Pentode, and combination tubes (Dual Triode, Triode/Pentode combination, etc).
2. Plate voltage Supply variable from 100V to 350V, for Gm Measurements. The voltage supply is current limited to 100mA for tube protection. This should be adjustable to help prevent damage to smaller tubes.
3. Plate ‘Current Source’ Supply variable from 0.1 to 100mA for Mu measurements.
4. Screen Voltage Supply adjustable from 100V to 350V current limited to 10mA for tube protection.
5. Grid Voltage Supply adjustable from 0V to –30V with current limit to 1mA for tube protection. No positive Grid drive? Are there tubes that require positive screen drive during testing?
6. Cathode directly connected to ground.
7. Voltage to Plate, Screen and Grid measured WRT ground and available for displayed.
8. DC Plate Current sensed across a 10 Ohm 1% (or better) sense resistor.
9. AC Plate Current sensed across a 10Ohm 1% (or better) sense resistor for determination of Gm.
10. AC Plate Voltage sensed directly at the plate with the Current Source supply driving it for determination of Mu.
11. DC Screen current measured across a 10Ohm (1% or better) sense resistor.
12. 100mV RMS 1KHz sine wave source for driving the grid to facilitate Gm and Mu measurements.
Comments please.
Those Mercury testers are pretty nice. I have one that I end up using a lot even though I have many Hickok's (AC 47,TV-7,539,600,750) Sencore, Simpson, some Tektronix internal test rigs and even a Tek 570 tube curve tracer. I've been collecting them since the 60's.
My favorite is a Weston OQ-3 Navy tester. A real antique with all the meters and a complete patch panel, in an oak "portable" case. But I don't actually use it much, more of a show piece.
I use the Mercury because it is quick and easy, small and light and is usually sitting somewhere handy. Did I mention it is light? If I care more, I will use the 539 (or 750) or if the tube won't work in the Mercury. If I really care I use the curve tracer. Then they go in the circuit for an actual test.
I use the Tek engineering setups to check out tubes I don't have any data for, ditto with the tracer.
The Hickok 600 is a good all around tester.
They all have their limitations. The RAT tester is a nice option because it can be built to best suite the tubes you intend to use. One of these days I might build one (or something like it) but with so many testers already it's hard to justify.
My favorite is a Weston OQ-3 Navy tester. A real antique with all the meters and a complete patch panel, in an oak "portable" case. But I don't actually use it much, more of a show piece.
I use the Mercury because it is quick and easy, small and light and is usually sitting somewhere handy. Did I mention it is light? If I care more, I will use the 539 (or 750) or if the tube won't work in the Mercury. If I really care I use the curve tracer. Then they go in the circuit for an actual test.
I use the Tek engineering setups to check out tubes I don't have any data for, ditto with the tracer.
The Hickok 600 is a good all around tester.
They all have their limitations. The RAT tester is a nice option because it can be built to best suite the tubes you intend to use. One of these days I might build one (or something like it) but with so many testers already it's hard to justify.
The best tube tester is the toy that smoking-amp uses in this thread to force us drooling: http://www.diyaudio.com/forums/tubes-valves/159892-nice-triode.html
The 570 is even nicer (although it is all tube inside, which is a plus and minus).
Not sure what smoking has set up for his 576, but the 570 has drop in panels for various tube sockets with patch leads (I have tons of socket panels for mine). You can switch back and forth between two tubes, for matching. They are a kick to play with. You just dial in the plate voltage and current settings, then set the grid step sweep and play.
But the 576 and 577's are much easier to find and smaller/lighter.
The 570 is a lot like the 575 except for tubes. You could get this 575 and hack it pretty easily.
TEKTRONIX 575 TRANSISTOR CURVE TRACER WITH CART & MORE - eBay (item 220536066138 end time Feb-04-10 08:06:47 PST)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.