T hi

Digital Tube tester compare.

  Model Manufacturer
SOFIA 20 Audiomatica, Italy
ROEHTEST, Kit version only

Helmut Weigel, Germany

AT1000 Amplitrex, USA
L3-3 with Agilent PC interface. DIY Project.

Color Scheme
The Winner
This needs improvement
This is unacceptable to me




2) Roe-Test
Hardware Version 8
Note, lower Versions may have lower specifications at some points. So when you buy a used one, check it's version.
3) Alladin-II
4) Amplitrex AT1000
Manufacturer Homepage
6) L3-3

Obsolete Product.

This is a KIT. You must build it yourself. Functionality exceptional high

Under construction.
Almost finished.

This is based on the Testers we bought in 2009 and 2010

This is a KIT. You must build it yourself. Functionality exceptional high. Very compact circuitry, it works with an old Dell Notebook power supply that you can buy on Ebay for 10€. So mains transformer is needed.
With my own build digital interface on it. Under construction. Not for sale here

* Example picture only, because you need to build this tester yourself;

5500 Deutschmark in 1995.
Below 100€ for the build guide on CD. 800€ for the electronic parts. The housing price depends a lot on what you want, but appr 200€ is the direction. So calculate with appr 1100 Euro.
  For a price of 2400 US$ in 2012, it was the only girl in town at that time. Price on Ronald's homepage, 250€ by Paypal. This includes all parts for the PCB + Software. You must buy yourself: A old Dell Notebook power supply , a case, and tube sockets.
Ebay prices from 250...400 Euro. Add the Agilent 20 Channel multimeter 34970 A with interface card. On Ebay for appr 500 Euro. This basically works. A PC Interface costs appr 100 Euro.
Repair Support

They offer repairs still after 25 years. Provided they can buy replacement parts needed.

Tester Manual

Repairs are done, and seem hardly needed.

Message from the president:

Dear Mr. van de Walle, as Mr. Bigi told you over the phone we do not supply schematics of our products. This a Company policy since 1985 and up to now we had only elements to enforce it.

Best regards,
Maurizio Jacchia. President Audiomatica Srl

(Received By email, 19.07.2011 16:06)

  not officially
Linux compatible
Probably with DOSEMU, but I never tried it.
Works with WINE
Not that I know
Not that I know
Windows compatible
All Versions from XP to W8
  All Versions from XP to W8
All Versions from XP to W8
16bit. Which is achieved by external D/A converters. So no analog Single Chip inputs. When you go to very low voltage, the Sofia has 100mV offset, which is can be a lot when you test tubes with low grid voltage. I am pretty sure this is a calibration issue. So the tester itself can not be blamed after 20 years.
12 Bit for V5 or higher. Though at small signal, the tester auto scales to more sensitivitry. This gives an artifical 16bit result.

10bit, and I can confirm it is really there and no offset problems. However it is not enough.

Also, it is presented better than it is. For instance 6SN7 tests in steps like this: 2550, 2650, 2750. The last two digits are always "50", so the last two digits are of no use. To my opinion a poor result.

22 Bits when I use Agilent 34970A.
Measurement range Anode
Absolutely AMAZING. It can do 0...700V DC, 250mA
Anode: 0-51 V, 0-306 V. With Optional Power Supply Card: 300-606 V. Each range 250mA, peak 300mA. Tube curves may 300mA.
0..300V @ 50mA


Below 75V not possible

0..400V at

0..250V @ 100mA

300V at lower current

Measurement range G1

0-5.1V, 0-63V. Each 5 mA.

This is not enough for large triodes like 300B

-17V….0.02V @ 20mA

This limites the tester to specific tubes only.
0-100V. 150V in Auto Bias.


This is not enough for large triodes like 300B

0-15V 0-150V
Measurement of grid (G1) current  
I had a 45 tube, which was tested with 32uA grid current by the AT1000. Then, on the AVO Mk4 (analog tester) it had only 4uA. I added a precision current meter in the grid circuit of the AT1000. Result: 3,7uA. That is 865% error for the AT1000. A wrong test test result.
With electronic nano-Ampere meter , which can even be calibrated with internal reference. Best I know.
Measurement range G2   0-306 V, 60 mA    
Measurement range Heater

Filament 0..24V 5A (Wow!) and Current measurement included.

Filament: 0-12,75V, 5A. 0-127 V 500mA.

14V @ 1,2A

12,6 Volts

PC and or stand alone
Only PC
Only PC

Both. PC only windows.

ALSO there is a large display in the machine, indicating many data.

Only PC
Shorts test
No. This is dangerous. one day you plug in a shorted tubes, or you connect something wrong, with those loose cables.
Any short tets, it can be done. Even has an electronic nano ampere mater for this.
Gas test
No, See above, because a gassy tube may short spontaneously.


Hazard Lamp. An LED costs 10 cents only. With 400Volt unprotected at the sockets, this is saving money at wrong place.
Sound Test
Hardware access to all tube connections
Yes, by hand switching
With full relay matrix
Only Plate, Cathode and 1st grid. Several tubes can not be tested, and need socket adapters.
Almost. I would say yes
Dust proof cover
Depends on yourself.
Measure Filament current
Up to 6.3A
Hardware Quality

Hardware is very good. No fan. What I like, there is almost no micro stuff in there. Even IC's. So if you damage something, you have a good chance to fix it yourself.

The "secret" schematic issue, I accept during normal production, but not if production is stopped. For me personally, I would never buy an new production This is my personal opinion. If you buy a new product and they refuse to give you the schematic along with it, for me personally this has an unhygienic taste.

Depends on yourself. Service is convenient, due to modular construction.
Very good, but the Fan is definitely too loud.
Absolutely superb
Software updates
No. Version 2.1 for DOS is available from the Audiomatica Website, but this is not an update, it is very old. Software is fully mature, does not need an update

You download the sofware for free.

Software is not compatible with very old models.

  Updates. What updates? The version I have here, is a collection of mistakes and errorz.
Curve tracing
Curve charts are drawn the way they are supposed to look, right away. Scaling can be done afterwards, even if STORED curves. Wow!!!! This is an EXCELLENT Software job.

Optics of the curves not as nice as the SOFIA. It has the "Excel Look" like Amplitrex has, but options and possibilities are the best of the best.

What is missing is the classical curves as you see in any datasheet. Yet is measures very interesting, and professional.

Very good. By default it draws only one curve. It needs a lot of trial and error before you have the curves drawn the way the should look. I am not enthusiastic about this. Also the "Excel look" of the curves is not nice.
Curve charts are drawn the way they are supposed to look, and great care was given, to make give curves look original shape and functionality as in old datasheets. The others can take this as an example. A nice improvement would be a way to plot them as vintage "datasheet" look alike curves.
Speed Once the test is started, the curves are plotted after just seconds. Dynamical and static data is derived afterwards from the curves. Excellent feature: When you press"start" too quickly, curve plotting is auto delayed until the tube heater appears warmed up. This is done by monitoring the heater current draw. (I checked it with a scope) Great software job.     The AT1000 is not very fast. Depending on what you do, the Sofia can be 10x faster, and produces higher resolution.
Review saved results

Another top class feature of this software. You can save the charts to disk, and you can change the scales afterwards. So you zoom in and out. Like test the tube from 0....600V. Another day, you can plot a curve from 0...200V only, it uses the original measurement.

Apart from that, you can also save it as bitmap.

Curves are not saved in graphic format, but internal rsults are saved. Like this, you can re-scale the curves later.

Apart from that, you can also save it as bitmap.


Not really. Main thing for me, the "test result" page and "tube curves" can not be saved electronically.
Tube parameters at multiple operating point

Wow! A most remarkable feature is used here. You can move a marker around in the curves, and you get all tube data at the specific operating point. Even in STORED curves you can do this later. So you use a 300B at 320V or at 480V, and that makes a LOT of difference. Amazing this part of the software works still under XP. Check here how to install that freeware.


Software Maintenance

One major bug, see below. For the rest, I found nothing that needed improvement. Such a result after 25 years is more than magnificent.
This is living software, suggestions and improvements are implemented in the next versions. Perfect attitude.

I reported many issues and small items. My impression is, Amplitrex is not interested in improvements. Don't think I never tried, but it was a waist of my time.


Software (1)

Pentode curves

Also Ultra Linear, or triode mode is possible
Also Ultra Linear, or triode mode is possible
It tests pentodes in triode mode. Problem Reported to Amplitrex, it seems technically possible, I have a working test version, but they never cared to follow up on it.

Software (2)

Diode Curves.


Officially the answer is No. I found a way to do it myself, even for twin diodes.


Software (3)

Grid voltage

for DHT tubes

Grid voltage is measured wrong for Directly Heated Tubes. The effective grid voltage the tube sees, in increased by half of the filament voltage. This was not compensated in the software. (Though this would have been very easy to do).

So when the tester draws a -1Volt line for a 300B, it is in reality the -3,5 Volt line. Same error is made with the normal"Tube Test" page it produces. With the Sofia, thinking of an improvement is 20 years too late of course.


DHT tubes are measured same as in the historical datasheets, when you use the function: "Simulate AC heating"



Grid voltage is measured wrong for Directly Heated Tubes. The effective grid voltage the tube sees, in increased by half of the filament voltage. This was not compensated in the software. (Though this would have been very easy to do).

So when the tester draws a -1Volt line for a 300B, it is in reality the -3,5 Volt line. Same error is made with the normal "Tube Test" page it produces. With the Amplitrex, I only report it here, for everbodies interest, but reporting it to Amplitrex has only resulted in a work-around advise.


Software (4)


This is an issue with the AT1000 only. In the stand alone mode, Fixed Bias, the transconductance is represented in percent (of something). Problem is transconductance (in percent!) is only meaningful at 100% plate current. However at Fixed bias, 100% plate current is a matter of coincidence.

Software (5)

Emission Test

You can simulate it with set ups.
You can simulate it with set ups.
Though the software is telling you "Emission" and even shows this in percent, it is actually not measuring Emission at all. Just measures plate current and calls this Emission. That's a real slip.
You can simulate it with self made cards easily, but you do need to understand how this is done, and also you need to know yourself how to make cards. This is not presented on a silver platter for free in the internet.
Matching (6)


Most excellent method used, which is a MASTER EXAMPLE of engineering quality. This method if HIGHLY RECOMMENDED to study, for every one who wants to make software for tube matching. So you can test 100 tubes, and the software tells you what tubes have CURVES that fit well to other tubes. Not just one or two test points, but full CURVES matching.

Helmut Weigel writes me, this is on his list, but a date is not known yet.
No. You have to decide for yourself if what you see is a good match or not. Specially since this tester has all data and options on board, and for what you pay, this is simply MISSING.
No. You have to decide for yourself if what you see is a good match or not, but you have to see this is a curve tracer in the first place.
Distortion Measurement
Yes, but to my opinion the method is not very good. Gives second and third harmonic Distortion Measurement, what maesurement?  
I think when you are very good, you might add it, with some method to bypass the bandwidth filter for Gm. The difference between filtered and unfiltered is the added noise, which is composite distortion. Not exteremely hard to do actually.
Quality Control (8)
So the tester tells you if a tube is good or bad.

So the tester tells you if a tube is good or bad.

Indication is like red/yellow/green meter.

No. With "quality control" I expect the tester to tell me if a tube is good or bad. Add to this the plain errors with Emission Test and Transconductance test.

So the tester tells you if a tube is good or bad.

Indication is like red/yellow/green meter.

Export of Test report
Yes. by hand or auto mode. Saves a bitmap
For the regular test reports, all you can do is print them.
The future of this software

The Microsoft Gods have decided not to allow direct access to the RS232 ports in all Windows Versions, meaning the Sofia can only run under DOS.

* However as MS obsoleted DOS, with the coming of W7, something nice took place: The open source community started to make DOS emulators. These all have the problem that Windows per definition will not allow direct access to the RS232 and LPT1 port, and besides, DOS expects an RS232 mouse on COM1, and nothing else. I see good progress with newer DOS emulators. Right now a 100% good solution does not exist. Thought t seems DOS BOX Megabuild Version may work. (Only this version and no other) Use an old DOS PC for your Sofia is the best.

As it also runs under Linux, we have an excellent reserve option here. "Green" score!

The software uses Microsoft internal modules, like MSCHART, etc, which is part of the microsoft Operating system. However these are very old microsoft modules, and I don't know what happens if one day, the Microsoft Gods decide, only newer versions are allowed to run. From how I know Microsoft, I see this as a potential risk myself. I hope I will be wrong.


Pro: No-compromise test range with excellent resolution. Highly professional software, with original and normal looking tube curves, ready to put in a datasheet (Example here, scroll down)

Contra: You can only install it when you understand MSDOS.

Pro: This products is alive! Each newer version fills this matrix with more green boxes. Incredible complete tube data base. Excellent email support.

When you are very technical and when you can build something, you should have this machine high on your whish list.

Contra: G1 is too low for the optional 600V Anode card. Appearanc of tube curves is a bit unuseual. And that is all "contra" I have.

Not released yet.

Pro: The good hardware.

Contra: Sofware full of bugs and errors, and no updates.

Pro: Very good hardware, low price, and very mature product. Easy to build, and needs not much time.

Contra: A very small one. RS232 ports, your PC may not have that. So this adds to the costs, and also installing uncertified, mostly Chinese RS232 cards, is getting harder with each newer version of Windows.

No intention to compare with the others. Regard it my personal project. I want to use it later for special tests, that I can program my own way, and simply do things the way I want.