[Milsurplus] The FR-38 Saga, and GF-11 diodes

Mon Dec 4 18:02:24 EST 2017

BobWell, since we are speculating on what the designers specified for these switching diodes, let's guess that the part number "212 GF-11" indicates that these particular GF-11 diodes were selected by test to withstand 212 volts reverse voltage.
Since diodes rated at 250 Vr can be had for 25 cents each, including postage, if bought from Mouser or 6 cents each if bought from Singapore, why would you not use these diodes?  They are certainly cheap enough.  If one enjoys replacing parts, well that is a whole other story and I would not stand in the way of the enjoyment one gets from working on old vacuum tube gear.
Once the knee in the reverse bias mode is passed, current rises exponentially.  True, if there is enough resistance in the circuit, the diode junction will not burn open since the resistance limits the reverse current.  But that does not mean that the PN junction was not damaged in the process.  Let's speculate that the spare GF-11 diodes were used to replace the diodes that slowly failed over the life of the counter.  Remember, we are talking about Germanium diodes and all of their problems that do not exist with silicon diodes.
We can also take a stab at why 4.5 volts was chosen as a test voltage.  Incoming Inspection just needs to know one point of the breakdown curve to know whether or not the diodes pass the breakdown requirement.  If they pass, fine; if not they are undamaged and can be returned for credit on the next order.
Just after I got out of the service, I worked at a company that 100% inspected 1% resistors, temperature cycled them over the full military temperature range and then re-inspected those resistors.  This process took days for orders of a 1000 resistors per lot.  If the resistors did not hold the value to whatever parts per million the test document specified, the whole lot was returned and the process repeated.  Spec is spec.

Well, most of the Christmas decorations are now up and the wife is happy.  Happy wife, happy life!Jim
      From: Bob kb8tq <kb8tq at n1k.org>
 To: Ray Fantini <RAFANTINI at salisbury.edu> 
Cc: Jim Whartenby <antqradio at sbcglobal.net>; Milsurplus <Milsurplus at mailman.qth.net>
 Sent: Monday, December 4, 2017 2:20 PM
 Subject: Re: [Milsurplus] The FR-38 Saga, and GF-11 diodes

Hi
Well …. if the test set uses 4.5V, that suggests that they were not real worried about 150V ….
One of the max voltages will (obviously) be when the “drive" tube is cutoff and the “gate” tube is cutoff by another input. Most of what I used to work on had interesting sets of bias resistors to get this and that done. You loose a bit of gain since the bias circuit is an attenuator. There still is plenty of drive. 
Transistors are a lot more “user friendly” for DC cascading than tubes. That was not always as obvious as it is today.  Dig into early transistor based computers and you will find all sortsof odd supply voltages running around. 
Bob

On Dec 4, 2017, at 2:50 PM, Ray Fantini <RAFANTINI at salisbury.edu> wrote:

#yiv7927754205 #yiv7927754205 -- _filtered #yiv7927754205 {font-family:Helvetica;panose-1:2 11 6 4 2 2 2 2 2 4;} _filtered #yiv7927754205 {panose-1:2 4 5 3 5 4 6 3 2 4;} _filtered #yiv7927754205 {font-family:Calibri;panose-1:2 15 5 2 2 2 4 3 2 4;}#yiv7927754205 #yiv7927754205 p.yiv7927754205MsoNormal, #yiv7927754205 li.yiv7927754205MsoNormal, #yiv7927754205 div.yiv7927754205MsoNormal {margin:0in;margin-bottom:.0001pt;font-size:12.0pt;}#yiv7927754205 a:link, #yiv7927754205 span.yiv7927754205MsoHyperlink {color:blue;text-decoration:underline;}#yiv7927754205 a:visited, #yiv7927754205 span.yiv7927754205MsoHyperlinkFollowed {color:purple;text-decoration:underline;}#yiv7927754205 span.yiv7927754205apple-converted-space {}#yiv7927754205 span.yiv7927754205EmailStyle18 {color:#1F497D;}#yiv7927754205 .yiv7927754205MsoChpDefault {font-size:10.0pt;} _filtered #yiv7927754205 {margin:1.0in 1.0in 1.0in 1.0in;}#yiv7927754205 div.yiv7927754205WordSection1 {}#yiv7927754205 The stages on the most part are binary flip flops that have diodes from two different DC busses that are routed to the plate circuts. One bust is +70 volts and the other +90 volts. The common B+ bus runs at +210 volts. Others are in the grid circuit and run to the -70 volt bus.Like you say they all appear to be used for forcing a tube to go into the high or low state and latching them there.Capacitance would be a huge issue, using a 10X oscilloscope probe is often enough capacitance to cause a stage not to work correctly coupling capacitors are in the 22 pF and 27pF so all this stuff is real high impedance in nature also. Nothing like solid state electronics.The GF-11 is the only type diode used in the counter, that along with a mysterious big blue block that appears to be a full wave bridge that’s across the incoming AC bus and feeds the DC blower motor.Another interesting foot note is the manual tells how to build a leakage current test set up that uses 4.5 volts, 100K resistor and a 50Ua meter to read leakage current. They have a chart that gives a Go/No Go value for the leakage current and instruct that you have to elevate the temperature of the diodes to get an accurate reading and this is the preferred method for testing and not to use an Ohm meter because of the differences in voltage used. I did not do this and just relied on looking at resistance with an analog meter on the 1 K scale and that worked for me. May build up the test set later but then back to the same problem of needing replacements.    Ray F/KA3EKH       From: milsurplus-bounces at mailman.qth.net [mailto:milsurplus-bounces at mailman.qth.net]On Behalf Of Bob kb8tq
Sent: Monday, December 04, 2017 1:11 PM
To: Jim Whartenby <antqradio at sbcglobal.net>
Cc: Milsurplus <Milsurplus at mailman.qth.net>
Subject: Re: [Milsurplus] The FR-38 Saga, and GF-11 diodes   Hi    Let’s back up a bit.     Most of this stuff is doing basic logic with diodes. Put maybe four diodes to the grid of a tube. Each one goes  back to a plate circuit of another stage. Pull any one of them “low” and the tube turns off (if set up one way) or pull any one of them “high” and the tube turns on (if set up the other way). A twenty volt grid swing is  likely plenty to turn a typical triode on or off.     If the cathode of the tube is roughly at ground, the grid will be running from -10 or -20 to around zero volts.  The preceding stage output is padded / biased down to be in this range. None of the diodes (in normal operation) see a lot of voltage. The grid bias *is* often high impedance. Leakage (or capacitive feedthrough) is a problem. They would probably go to higher voltage rated parts to get lower leakage.     Once you have silicon, leakage becomes much less of an issue. A big junction with silicon still has a lot of capacitance. That’s one of the things that makes a diode slow. There are a few other things as well.     Are all circuits done the same way? Of course not. Without the schematic, there’s not much of a way to *know*  what HP did in this case. They didn’t have a lot of “prior art” to look at when this counter was designed. Solid state  diode logic was very much a new and scary technique back in the late 1940’s.     Best to dig into the schematic and do some scope measurements. See what the diodes are getting hit with and pick something according to what the circuit really is doing. If it’s always in the < 30V range, go with a > 60V rated part. If it actually does see more than 100V, by all means, don’t use anything with less than a 200V rating. Keeping down the rating (if possible) gives you a bigger selection of fast diodes to pick between ….    Bob    

On Dec 4, 2017, at 11:49 AM, Jim Whartenby <antqradio at sbcglobal.net> wrote:    Ray the 1N914 has a 75 Vr rating.  See if you can find the 1N645 which has a 225 Vr rating, I would think that this would work well with tube circuits and voltages. Jim    From: Ray Fantini <RAFANTINI at salisbury.edu>
To: "mrca at mailman.qth.net" <mrca at mailman.qth.net>; Milsurplus <Milsurplus at mailman.qth.net> 
Sent: Monday, December 4, 2017 8:00 AM
Subject: [Milsurplus] The FR-38 Saga, and GF-11 diodes    Ended up using just some generic 1N914 diodes that I had laying around the shop. Had only two original diodes that were way under the minimum reverse resistance per the manual specification of 75K Most of the Germanium diodes if not all have some leakage and would be good to replace most of them at some point in the future but the replacement of the two leaky ones got my gate trigger MMV up and running. The next problem was the decade divider that divides the signal from the time base was only dividing by two. That’s mounted on a sub assembly and is removable so after removing, checking the tubes and cleaning it up somehow it magically started working again. Suspect that there may be a bad 0.01 cathode bypass that’s causing it to run as a MMV and not a flip flop and may go ahead and change that while it is out. With the gate circuts now all working and after completing a time base calibration late last night I was able to do a successful self-test of the 100KC and 10 MC internal reference oscillators. All of this was with the input section removed but have seen it generate output so going to assume that there are no issues with that. The next step will be putting everything back together and all the covers on and seeing if it still works then. Working on vacuum tube digital electronics has been a learning experience and armed with a good dual trace scope and service manual not as daunting a task as I assumed it would be when starting but have to say that without proper documentation and the excellent service manual that not only included schematics, waveforms and most important theory of operations would not have been able to do it. Also thanks to everyone who helped out with the diode question. Think what I will do on that front is maybe next year at Dayton will stop of at Midwest Electronic Surplus and see what they have in a bunch of identical small signal switching diodes and try the dreaded shotgun approach to try to keep all the diodes the same type.   Ray F/KA3EKH ______________________________________________________________
Milsurplus mailing list
Home: http://mailman.qth.net/mailman/listinfo/milsurplus
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Milsurplus at mailman.qth.net

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html ______________________________________________________________
Milsurplus mailing list
Home: http://mailman.qth.net/mailman/listinfo/milsurplus
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:Milsurplus at mailman.qth.net

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.qth.net/pipermail/milsurplus/attachments/20171204/d4f3cea4/attachment-0001.html>