It's not what it sounds like, and "Yes" I did plug it in.

I'm trying to build some calibration devices for . . . oddly enough, a Vacuum tube tester. (Yes. Vacuum tubes are still in use.)

The first problem is that the tester uses about 20 D'Arsonval meters. Not easy to find or get data sheets for. Some of these meters are 50 years old, so we don't want to blow them up, but they do need to be calibrated, daily. Unfortunately, that means that we need a variable current source that works between 100 pA and 500 mA. I know that you have all heard that you shouldn't test a D'Arsonval with a DMM. There's a good reason for that.

What I decided to do was buy some very expensive IC current limiters. But what do I do if they don't work as intended? I've been looking at low current transistors, of several types, but I haven't found any schematics that show something that appears to be worth pursuing. Sure, I could design a circuit to make a transistor that would do the job, but there are stability concerns. I don't understand the design parameters of stability circuitry that well.

Second problem is that I need a 0.5 VDC reference. Hmmm. Isn't that lower than the Vbeo? Yep. So where do I start looking? And voltage dividers are not the answer.

Do any of you work in this voltage/current range, or know someone else who does?

I've only been working on this for a month. A lot of that time was spent on Google. I haven't been able to even find a lead. It's frustrating.

In situations like this, in the past, it seems like the major problem was that I was barking up the wrong tree, but it's the only tree I can see.

I need new input.

BTW, I got perma-banned from another electronics forum for pursuing this search. Hope that doesn't happen here.

 

Depends on what you want to spend and what you consider expensive? The Fluke 725 is a nice multifunction process calibrator which will do everything you want and more. When you say:

Second problem is that I need a 0.5 VDC reference. Hmmm. Isn't that lower than the Vbeo? Yep. So where do I start looking? And voltage dividers are not the answer.

DC reference implies no real current sourcing just a very accurate DC source. The above mentioned Fluke 725 will do that. You haven't defined the accuracy uncertainty you need or want. As to the higher currents? Look towards a used Fluke calibration source. Look for a Fluke 5700A and 5720A but you are getting into some pretty expensive calibration standards. Just a matter of what uncertainties you can live with and what you want to pay. I don't see where a +/- 2% or 3% FS meter movement needs much in the way of precision. If you could find an old Fluke 760 on the used test equipment market it would be ideal. Just be aware good calibration standards equipment isn't cheap and there is a matter of maintaining it with periodic calibration and certification as required or needed.

Yeah, vacuum tube testers are still in use and a few companies have taken to marketing new design vacuum tube testers. While I don't see what a vacuum tube tester will tell me that I can't see looking at a tube in circuit whatever trips their trigger I guess.

Ron

 

For a 500 mv voltage reference, I would start with an easy-to-get and adequately stable voltage reference; perhaps the 2.5 volt ±1% TL431A would be sufficient, and then use a resistive divider to get down to 500 mv. If your reference needs to have a low impedance you can follow it with a suitable opamp configured as a voltage follower.

Robert Pease published an enlightening article on the subject of measuring femtoampers, maybe you can pick up some tips from it. http://electronicdesign.com/test-amp-measurement/whats-all-femtoampere-stuff-anyhow

You could probably use a Burr-Brown OPA655 from Texas Instruments (5 picoampers -and I know both Philips and Analog Devices have made op amps with lower bias currents) , to make a Howland current source (www.ti.com/lit/an/snoa474a/snoa474a.pdf).

The easiest way of making the 50 picoamps is probably by using a low voltage source and a huge resistor; 50 mv and 1 gigohm comes to mind for the 50 picoamps, higher voltages for higher currents. Be sure so clean your fingerprints (and everything else) from the board before calibration, and after researching this further you might decide to use teflon as your substrate.

Additionally, you might have to connect and disconnect your higher-current current sources using a relay to keep leakage from that source from affecting your 50 picoamp current source.

Hope this helps some.

 

Depends on what you want to spend and what you consider expensive? The Fluke 725 is a nice multifunction process calibrator which will do everything you want and more. When you say:



DC reference implies no real current sourcing just a very accurate DC source. The above mentioned Fluke 725 will do that. You haven't defined the accuracy uncertainty you need or want. As to the higher currents? Look towards a used Fluke calibration source. Look for a Fluke 5700A and 5720A but you are getting into some pretty expensive calibration standards. Just a matter of what uncertainties you can live with and what you want to pay. I don't see where a +/- 2% or 3% FS meter movement needs much in the way of precision. If you could find an old Fluke 760 on the used test equipment market it would be ideal. Just be aware good calibration standards equipment isn't cheap and there is a matter of maintaining it with periodic calibration and certification as required or needed.

Yeah, vacuum tube testers are still in use and a few companies have taken to marketing new design vacuum tube testers. While I don't see what a vacuum tube tester will tell me that I can't see looking at a tube in circuit whatever trips their trigger I guess.

Ron

Eeeew. I would love to have one of those! Unfortunately, I make $1500 per month. Not sure about the other guy. What a beautiful piece of equipment.

For accuracy, we have agreed on +/- 0.5%. Less would be better. And TYVM for the information on DC references. I never really thought about that, or maybe just forgot about it 40 years ago. A mind is a terrible thing to lose.

The problem with the meters is that we are buying a lot used, and they don't come with data sheets. It's embarrassing to hook up a brand new, slightly used, meter and have it kick once, then die. Yes, that has happened to me more than once. Until I can come with a decent protection circuit, I'm not going to hook up any pico-ammeters at all.

The thing is that the testing takes days some times. As I understand it, the meters are considerably more accurate after the circuit stabilizes, especially if calibrated more than once a week.

Glad you mentioned vacuum tube testers. I got into this because I used to rebuild them to better than factory specs. I even replaced all of the wiring in a TV-7D once, during what was supposed to be a cap/resistor replacement.

The thing is that we are looking at a niche market. With a matched pair of of 12AX7s going for $3000 we should be able to buy in bulk. Buy 100 of them for $1500 and, if we find even one matched pair, you have made a 50% profit, plus however many good singles there are.

The problem is credibility. The number of audiophiles who can afford these is small, and many of them know each other. So, we sell one substandard tube, or worse, pair, and we will never sell another one. Even a Hickok, or a TV-7 could put us out of business.

We are at the beginning of the design process. We have decided what we need, and have some ideas for some of the equipment we need, but there are a lot of holes. After we get a working model, phase 2 is to start making it modular and automating it. We can't afford to do either of those until we sell a few tubes.

So this really is a project. Fortunately, or unfortunately, I received my electronics training in the US Marines, starting in 1968, and ending in 1987. The fortunate part was that, when I started in 1968, the Marines used tube radios discarded by the Air Force in 1957. RIFed after Desert Storm, I spent 15 years as a Process Control Systems Consultant, and met my business partner, also a PCSC. He's Russian and his father is, or was, the manager of the largest bank in Moscow.

I'm sorry about running on, but I didn't explain myself very well on another electronics web site and go perma banned. Needless to say, I'm still disturbed about that.

 

Alas, the old TV7 series, I remember them well, yes, I am that old and worked on them during my military years. Your kidding? I can't believe the cost of matched pair tubes. During my early Tektronix scope years (the 545B comes to mind we would buy 12AX7 matched pairs from Tektronix by the truck loads. You have got to be kidding:

I received my electronics training in the US Marines, starting in 1968, and ending in 1987.

Been there and done that. USMC 69 to 79 when I went DoD and worked for the Navy.
Thinking about this with knowing more now than before I would be thinking of sort of rolling your own test equipment. You likely know this but on the used old panel meters sometimes they have in small print on the face the full scale current.

Semper Fi
Ron

 

For a 500 mv voltage reference, I would start with an easy-to-get and adequately stable voltage reference; perhaps the 2.5 volt ±1% TL431A would be sufficient, and then use a resistive divider to get down to 500 mv. If your reference needs to have a low impedance you can follow it with a suitable opamp configured as a voltage follower.

Robert Pease published an enlightening article on the subject of measuring femtoampers, maybe you can pick up some tips from it. http://electronicdesign.com/test-amp-measurement/whats-all-femtoampere-stuff-anyhow

You could probably use a Burr-Brown OPA655 from Texas Instruments (5 picoampers -and I know both Philips and Analog Devices have made op amps with lower bias currents) , to make a Howland current source (www.ti.com/lit/an/snoa474a/snoa474a.pdf).

The easiest way of making the 50 picoamps is probably by using a low voltage source and a huge resistor; 50 mv and 1 gigohm comes to mind for the 50 picoamps, higher voltages for higher currents. Be sure so clean your fingerprints (and everything else) from the board before calibration, and after researching this further you might decide to use teflon as your substrate.

Additionally, you might have to connect and disconnect your higher-current current sources using a relay to keep leakage from that source from affecting your 50 picoamp current source.

Hope this helps some.

I've been looking at a 1.24 VDC reference on a chip. Found one I liked, but they are $15 each. They might well be worth it though, lots of protective circuitry and stability. I just don't like the voltage divider idea. It would probably work if I put 10 x 1%ers in parallel.

Will read all the links.

Teflon substrate? Build the circuit on Teflon? I usually do Brooklyn Style bread boarding. What an interesting reversal? TYVM. That might fix a few problems I've been having at high frequencies with other projects. Might that cause capacitance at high frequencies?

Howland current source? Will look it up.

Good point about the relay. I hadn't thought of that because I never considered what I needed to have powered up during calibration. I guess I was kind of thinking of always having everything except the reference circuits shut off when the calibration was being this done.

This site is great. Thank you all very much.

 

Alas, the old TV7 series, I remember them well, yes, I am that old and worked on them during my military years. Your kidding? I can't believe the cost of matched pair tubes. During my early Tektronix scope years (the 545B comes to mind we would buy 12AX7 matched pairs from Tektronix by the truck loads. You have got to be kidding:

Been there and done that. USMC 69 to 79 when I went DoD and worked for the Navy.
Thinking about this with knowing more now than before I would be thinking of sort of rolling your own test equipment. You likely know this but on the used old panel meters sometimes they have in small print on the face the full scale current.

Semper Fi
Ron

What is it now? 240 years of tradition, undisturbed by progress?

545B? I never heard of an oscilloscope before I got out of the Marines the first time. Not kidding. Did you go through school at San Diego, or did you go to Navy schools?

It's like I said, it's a small niche, but, as far as I know, the Russians are the only ones producing vacuum tubes any more, and they don't make a lot of different types. 12AX7s are as scarce as hen's teeth. Audiophiles say they make a huge difference. Fortunately I spent too many years standing next to cannons that were firing to be able to tell the difference except in the tube stats.

Frankly, we are lucky if a meter has numbers on the scale, and a few have had the scale sheets removed completely.

BTW, a TV-7, especially a D model, is worth more than $2000 now if anything lights up when it is turned on. But jeez. Half of it's weight is wire. That repair job took me more than 2 months.

If I had all the money in the world, I'd like to fix one to use sensors, microprocessors and fiber optics. WTH am I talking about? It would probably cost more than a Ferrari, and be heavier than the original version.

Great to hear from you.

 

Ooops. Para 3 may need a little explanation.

Audiophiles love 12AX7s, but they also like volume. Why not rock your neighbor's world as well as yours?

So the 12AX7s have a life expectancy slightly greater than a flashbulb. The other main use is in guitar amplifiers, and you know that, "If it isn't loud, it isn't rock and roll." (George Thorogood)

Unfortunately, that's where all of the 12AX7s have gone. I've heard that people are having to make do with 12AU7s now. So most of those are gone, too.

Maybe my buddy could get a Russian plant to start building 12AX7s again, but their quality control is still a little iffy.

Well, I don't have a radio, telephone, or TV, so what does it matter to me?

 

What is it now? 240 years of tradition, undisturbed by progress?

545B? I never heard of an oscilloscope before I got out of the Marines the first time. Not kidding. Did you go through school at San Diego, or did you go to Navy schools?

It's like I said, it's a small niche, but, as far as I know, the Russians are the only ones producing vacuum tubes any more, and they don't make a lot of different types. 12AX7s are as scarce as hen's teeth. Audiophiles say they make a huge difference. Fortunately I spent too many years standing next to cannons that were firing to be able to tell the difference except in the tube stats.

Frankly, we are lucky if a meter has numbers on the scale, and a few have had the scale sheets removed completely.

BTW, a TV-7, especially a D model, is worth more than $2000 now if anything lights up when it is turned on. But jeez. Half of it's weight is wire. That repair job took me more than 2 months.

If I had all the money in the world, I'd like to fix one to use sensors, microprocessors and fiber optics. WTH am I talking about? It would probably cost more than a Ferrari, and be heavier than the original version.

Great to hear from you.

<OFF TOPIC>
Boot was MCRD San Diego then after ITR in Pendleton on to NAS Memphis (Naval Air Technical Training Command Millington TN) for about 6 months of basic electronics. Nothing new as I was into ham radio in '63 at 13 years old. Then Cherry Point till I ended up in Nam in '72 (Started in Iwa Kuni and was with MAG 15 which deployed back to Vietnam. Left Nam in 72 and Recruiter School back to MCRD. Ended up 3 years Cleveland, Ohio as a recruiter. Enlisted my own sister! Left Cleveland in '75 and ended up at the US Navy Fleet Calibration school NAS North Island, San Diego CA where I was the senior instructor and ran the school, sweet duty. That led to my leaving the Corps in '78 or so and going with DoD through my old command. My dad was USMC WWII and the sister I enlisted ended up at Parris Island SC as a Woman Marine DI, she later married a Marine down there, another drill instructor. He eventually went LDO and retired at 20 years as a Captain in charge of counter intelligence for HMX-1. No shortage of Marine blood in my clan. Saw your comment:

Teflon substrate? Build the circuit on Teflon? I usually do Brooklyn Style bread boarding. What an interesting reversal? TYVM. That might fix a few problems I've been having at high frequencies with other projects. Might that cause capacitance at high frequencies?

I was born in Brooklyn NY February 7th 1950 and grew up in NY. Go figure.

</OFF TOPIC>

Ron

 

Selecting matched pairs is a different problem (and somewhat easier) than quantifying exact parameters.

I am not a vacuum tube expert -- so I cannot help you with specifics -- but, select the parameters you wish to match. Then, drive each DUT simultaneously with exactly the same current or voltage (does not need to be precise, just equal), and look for the differences. A bridge-type circuit and/or differential amps are excellent for tasks like this.

 

Selecting matched pairs is a different problem (and somewhat easier) than quantifying exact parameters.

I am not a vacuum tube expert -- so I cannot help you with specifics -- but, select the parameters you wish to match. Then, drive each DUT simultaneously with exactly the same current or voltage (does not need to be precise, just equal), and look for the differences. A bridge-type circuit and/or differential amps are excellent for tasks like this.

That part we have pretty much figured out, out except neither of us really has the hearing to figure out what 'mellow' is in terms of electronic parameters. I would really love to hear an explanation from someone who has a full range of hearing. You see, my buddy also spent too long standing next to cannons.

We believe in 'mellow' but really don't know what it is. It is amazingly hard to find an explanation. Google it and you get a lot of links to audio files. We still have the same problem. If we listen to the audio files we can't tell the difference.

We are looking for an actual parameter, that can be measured electronically, but we know we may never find one. Could it be a combination of parameters?

Your para 2 is what we plan to do to check matching. After a tube is tested, it's parameters are put into a database. If a match is found, we start the 'matched pair' test. We run the sockets in parallel to get the same voltages to all elements, with the same input, then check the output data.

A bridge would be lovely, if we knew what ' mellow' meant.

I find that is is a very interesting project no matter how it turns out. The number of parameters boggled my mind, and I've worked with tubes a lot.

TYVM for your input.

 

My guess is "mellow" equals "distortion". But you can put 20 different audiophiles in front of an A/B switch and get 20 different answers -- and that's with only two choices.

You are thinking with the wrong hat. Put your used-car-salesman hat on:

Step 1: Call yourself an audio expert, audiophile, whatever -- insist that anyone who disagrees with you is an idiot;
Step 2: Put two tubes in a box -- call them matched
Step 3: Sell them as "Superior Matched Tubes" at $6000 a pair.

90% won't be able to tell the difference, and the remaining 10% will assume they must be wrong because, well, $6000.

 

Oh, I forgot Step 1a:

Step 1a: Make up whatever specs you want -- if you can't measure it, neither can they.

 

<OFF TOPIC>
Boot was MCRD San Diego then after ITR in Pendleton on to NAS Memphis (Naval Air Technical Training Command Millington TN) for about 6 months of basic electronics. Nothing new as I was into ham radio in '63 at 13 years old. Then Cherry Point till I ended up in Nam in '72 (Started in Iwa Kuni and was with MAG 15 which deployed back to Vietnam. Left Nam in 72 and Recruiter School back to MCRD. Ended up 3 years Cleveland, Ohio as a recruiter. Enlisted my own sister! Left Cleveland in '75 and ended up at the US Navy Fleet Calibration school NAS North Island, San Diego CA where I was the senior instructor and ran the school, sweet duty. That led to my leaving the Corps in '78 or so and going with DoD through my old command. My dad was USMC WWII and the sister I enlisted ended up at Parris Island SC as a Woman Marine DI, she later married a Marine down there, another drill instructor. He eventually went LDO and retired at 20 years as a Captain in charge of counter intelligence for HMX-1. No shortage of Marine blood in my clan. Saw your comment:



I was born in Brooklyn NY February 7th 1950 and grew up in NY. Go figure.

</OFF TOPIC>

Ron

WAY OFF TOPIC

Same here until Memphis. I asked my DI where I would go after boot. I was hoping for 2nd Tanks. But he told me that I could see the place I was going from where we were. ?? ITR and then San Diego again for my first year of electronics training. BES, RFS, and ARRC. I didn't even know that there was a NAS option until I went to Crypto Repair at Miramar (??) The first day we were asked to introduce ourselves. I was at the end of the end of the last row. Had trouble getting into the vault.) Everyone started off with "A school at . . ., B school at . . ." I was thinking, "What the hell are they talking about?" So my turn came and I said I had gone to BES, RFS, etc., at MCRD San Diego. They kicked me out of the vault.

The next day I was back, but that was the first time I realized Air to Ground is NOT the same as Ground to Air.

Cherry Pit? I spent years there with H&HS-28, TACC. Except for a year with H&HS-18 TACC on Okinawa, and a year at the Tech Course.

Other Marines in my family? None. I was given the option to enlist instead of going to jail. I tried the Navy. They said, "No, but the Marines will rake anyone." ( This was during Tet)

Semper Fi.

 

You just hijacked your own thread...

 

Oh, I forgot Step 1a:

Step 1a: Make up whatever specs you want -- if you can't measure it, neither can they.

Ah, I beg to differ. They say they can hear it. Maybe I should start working on an ear meter. Wait. That's already been done.

Brings up an interesting point, though. Have you ever seen a reliable automatic volume control? The only thing I've seen that was even in the right hemisphere was called a 'Raysistor', used in TYA-11 van 5 channel audio amps.

The best description I can come up with is that it's a little like an opto-coupler that controls the resistance of a resistor. So rare that they had individual serial numbers.

 

You just hijacked your own thread...

Mea maxima culpa. Please don't ban me.

 

Step 2: Put two tubes in a box -- call them matched

With the tubes fetching such big money, what's the betting some enterprising scamsters aren't already doing that? .

 

With the tubes fetching such big money, what's the betting some enterprising scamsters aren't already doing that? .

About the same as the odds that anyone would notice...

 

Ah, I beg to differ. They say they can hear it...

See Step 1.