|
Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets. |
|
Thread Tools |
18th Mar 2018, 9:17 am | #41 | |
Hexode
Join Date: Sep 2003
Location: Cheltenham, Gloucestershire, UK.
Posts: 298
|
Re: Testing unknown transistors for max frequency response
Quote:
I have a reasonable range of Ge devices you can borrow to do some testing - and I'm in Cheltenham, so they're readily accessible. dave
__________________
Dave Teague Last edited by dave_n_t; 18th Mar 2018 at 9:23 am. |
|
19th Mar 2018, 10:51 pm | #42 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
Hi Dave
Thanks for the offer of the Ge transistors. However, I'm a bit worried that I'll damage them during my experiments and I've already ordered some online. They seem really expensive so I'm doubly concerned that I'll probably pop a few of them along the way. I might visit a local car boot to see if there are any old/damaged/cheap transistor radios that I could borrow some more Ge transistors from. I'll ask at work as well. But I should be getting several different ones arriving in the coming days. I also really need to make an external bias tee for my VNA that goes down to a really low frequency. I might use one of the company 8753ES VNAs to take the s parameters because these VNAs work down to 30kHz. Mine only goes down to 300kHz.
__________________
Regards, Jeremy G0HZU |
19th Mar 2018, 10:58 pm | #43 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
Quote:
__________________
Regards, Jeremy G0HZU |
|
19th Mar 2018, 11:38 pm | #44 |
Nonode
Join Date: Dec 2012
Location: Kirk Michael, Isle of Man
Posts: 2,350
|
Re: Testing unknown transistors for max frequency response
I had a look thro my "higher end" Ge transistors this morning:--
AF279 AF280 AFZ12 (good qty) ASY27 ASZ21 ( " " ) If you want a few, just say so. I would not want any back. Les ps (Argus) I have a few OC35 here, maybe time to think of a 6v RB107 for one of the 'bikes. |
21st Mar 2018, 6:15 pm | #45 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
Thanks again to dave and also to Les for the offer of parts. For now I've got hold of some common Ge transistors and I'll see how I get on with these.
If the results are good then I guess it's a case of testing different types so that would be the best time to ask if I could have some more to play with. The first one out of the bag is a classic AF117 PNP. This is old but unused and it has 4 legs. I know about the whisker issue but is there anything I should know before trying to test it on a meter for the junctions and also for current gain at DC? I'm going to use my Maplin Gold DMM on the diode checker setting and this puts out 2.3V and can deliver about 0.7mA into a short. My little old Altai analogue meter has a beta tester built in and the idea is that it biases the base with something like a 220k resistor and it has three external croc leads. I think it tests the part at 3V and the idea is you fit/remove the base croc to look for beta/leakage. Presumably this will be OK with a Ge part but are there any classic gotchas when testing these devices? I'll practise on a BC556 Si BJT to make sure the old Altai meter still works but I only have one AF117 and I don't want to kill it
__________________
Regards, Jeremy G0HZU |
21st Mar 2018, 7:06 pm | #46 |
No Longer a Member
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
|
Re: Testing unknown transistors for max frequency response
Thanks for that kind remark. When I work with metals I always keep the surfaces heavily protected with layers of plastic tape to prevent any scratches to the surfaces from filings, tools & clamps and/or other debris. It is especially important if there is some existing coating, paint or anodizing, but its better to always. And it is worth smoothing cut edges and hole edges and removing file marks etc with 600 to 1000 grade paper to get a good finish, even in places that are not visible, as it gives a more pleasing result in the end.
With the AF117, usually it is obvious if they are affected by whiskers with either inter electrode shorts or shorts to the case of the e,c or b wires. There was some advice about not testing germanium signal/RF transistors on some meters as the currents might exceed the b-e junction's ratings or perhaps even zener the junction. But in any test circuit running from say 9 to 12V with a collector current in the range of 0.5 to 2mA and a base current of less than 1/10 of that I don't think there would ever be any damage. Check the spec of the individual types. One of the main issues is not to heat them up excessively while soldering their wires so use a small clamp on the lead to extract the heat while soldering helps, or a socket instead of soldering, solves the issue. |
21st Mar 2018, 10:05 pm | #47 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
Thanks. I'll try and use tweezers as a heat shunt when soldering. I think I'll go for broke and test it in the s2p jig first.
Quote:
Would that cover most typical uses? Each s2p file will be huge because I'll probably test at 0.3 - 500MHz with 1601 steps. I tested the BC547B across 0.3 - 1000MHz with 3201 steps. I also test at very low drive power to prevent any compression. Even a tiny amount of compression will obviously spoil the data. I think I tested the BC547B at -35dBm and that seemed OK.
__________________
Regards, Jeremy G0HZU |
|
21st Mar 2018, 11:49 pm | #48 |
No Longer a Member
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
|
Re: Testing unknown transistors for max frequency response
I think that would be fine, only the exceptional germanium types used for UHF tV tuners would go any higher than that. Many types would be fine tested from 0.3 to 100MHz or maybe 200MHz, certainly to see if they would make good replacements for AF11x. My data book says the Ft of the AF117 is 75 MHz.
|
22nd Mar 2018, 2:20 am | #49 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
OK, I've done a load of S2p measurements on the AF117 this evening. It looks like the Ft crosses zero at 93MHz at 1mA 5V. This seems a bit high?
However, if I plot beta as below and extrapolate from 2MHz then beta is 39*2 = 78MHz. Either way it looks like my AF117 is working OK. The next question is how/where do I upload all my s2p files? There are 25 s2p files each of 400kb size. See below for the folder contents and the filename shows the Vce and Ic in each case. I forgot to shrink the dataset size so these are 0.3-1000MHz with 3201 steps (sorry). Note: My first quick play with the s2p data shows that this AF117 could probably be configured to oscillate up at 350MHz when set at 10V 3mA if I look at the negative resistance looking back into the collector.
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 22nd Mar 2018 at 2:35 am. |
23rd Mar 2018, 2:48 am | #50 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
I haven't had much chance to play with the S2p data but here is a plot of GMAX for the AF117 when biased at 7Vce and 2mA.
This shows the max power gain a designer can hope for from this device at this operating point if they also want the stability factor K to be >1 at the chosen frequency. So this is a very powerful graph. It shows that: At 10.7MHz you can expect to get up to about 29dB gain with K >1 At 21.4MHz you can expect to get up to about 23dB gain with K >1 At 28MHz you can expect to get up to about 20dB gain with K >1 At 50MHz you can expect to get up to about 15dB gain with K >1 At 70MHz you can expect to get up to about 12.5dB gain with K >1 Obviously if you stray away from 7Vce and 2mA Ic the data will change. This info only applies at this operating point and you have to match the device input and output very carefully to get close to this theoretical maximum power gain and you have to keep an eye on the K factor in case it dips below 1. This requirement for matching will mean that you can only realistically achieve these power gain levels with a narrowband (tuned and matched) amplifier at each of these frequency bands. Hope this is useful!
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 23rd Mar 2018 at 3:03 am. |
23rd Mar 2018, 3:28 am | #51 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
Here's GMAX at 1mA and 6Vce. The gain figures are obviously going to be lower at 1mA instead of 2mA but 6Vce and 1mA Ic appears on a couple of datasheets for this device. The Ft at this point is 92MHz according to my s2p data. This is where beta crosses zero. However, if I extrapolate from 2MHz then Ft is more like 82MHz at this operating point.
At 7Vce and 2mA Ic the Ft is about 123MHz.
__________________
Regards, Jeremy G0HZU |
23rd Mar 2018, 3:59 am | #52 |
No Longer a Member
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
|
Re: Testing unknown transistors for max frequency response
Jeremy,
Very cool test results. They (AF117) were pretty impressive transistors for their day and your data appears to agree. Some radios that used them were dual MW/SW radios , some went to 30MHz. The transistors were known also to be good in 10.7MHz IF amplifiers and of course fantastic in 455KHz ones and in those requiring no neutralization either, because of the low C-B feedback capacitance. I think the AF178 has about 1/2 the feedback capacitance of the AF117. Hugo. |
23rd Mar 2018, 4:13 am | #53 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
Thanks, this does look to be a fairly capable transistor. To show how valid my s2p models are I used the s2p model at 7V and 2mA to design an RF amplifier at 70MHz using Genesys. This involves adding a lumped L match at the output and looking at the input match. You can see this LC matching network in the simulator schematic in the second image below. The PNP transistor symbol next to it is really my huge s2p data file. This data file is an s2p model and obviously doesn't need a power supply or any DC biasing so this is why the circuit may look strange without a base bias resistor or a power supply. It isn't needed for the s2p model. The FMAX is 12.7dB here but I didn't bother to match the input. I only matched the output. The input match is already reasonable but this causes me to lose a fraction of a dB in gain. So I only see about 12dB gain at 70MHz instead of the theoretical 12.7dB.
I then built the 70MHz amplifier with the real AF117 and compared S11 S21 S12 S22 for the simulator version (using the s2p data) and the real 70MHz amplifier measured on the VNA after calibrating it with a snazzy Agilent Ecal module. See below for the results. There are 8 traces on this graph but it only looks like 4 traces at first glance. The agreement is very good With this s2p data it should be possible to do some fairly serious design work with these transistors. I'm probably not the first person to take accurate s2p data for the AF117 but maybe nobody has done it over so many operating points.
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 23rd Mar 2018 at 4:42 am. |
23rd Mar 2018, 4:27 am | #54 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
Quote:
__________________
Regards, Jeremy G0HZU |
|
23rd Mar 2018, 5:14 am | #55 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,894
|
Re: Testing unknown transistors for max frequency response
Those data files are very interesting. I don't think anyone has built thorough s-parameter stacks for the classic germanium devices before. It really does show what they were capable of. The question remains of is there anywhere on the web where these files could be put to make them available to anyone interested?
I suppose vintage radio people might be the only ones interested at this level. The dominant market for old germanium types seems to be guitar pedal makers, at least they seem to set the market prices. You'd have to transform from the four 2-port s parameters into 'tone', 'sustain', 'expression', and 'gain in record sales' to interest them. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
23rd Mar 2018, 7:39 am | #56 | |
No Longer a Member
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
|
Re: Testing unknown transistors for max frequency response
Quote:
That will be very interesting. I have long advocated the use of the AF178 as a replacement for the AF117 and OC170/1 transistors, especially in communications radios like the EC-10. One of the things that attracted me to the AF178 was the large case which was a better match physically for the AF11x plus the improved Ft and lower C-B capacitance. The AF185 I think also makes a great replacement for AF11x too, but they became very hard to get. I remember when I saw the first one in an electronic shop in NZ in the 1970's and was gobsmacked by it. There are plenty of AF178's still out there as they were made for TV and much of the stock wasn't used up. I had been using these for many years here in AU for all of my AF117/OC171 replacements in various radios. I initially suggested the use of the AF178 over a year or so ago on this forum and it has caught on to the extent that some other members have tried them and found them good for the task. So your tests on them will be interesting. I think they will be pretty hard to beat as an AF11x replacement. I believe that all AF11x should be replaced because the tin whisker problem appears inevitable for these. Hugo. |
|
23rd Mar 2018, 4:08 pm | #57 | |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
OK I'll try and take some s2p data for the AF178 later today or tomorrow. However, this is a much faster part and really deserves to be tested in an s2p fixture optimised for its package. I can mill one of these fairly easily but it won't be for a while. So the first s2p data for the AF178 will be measured in the existing LF BJT jig that means the legs will be cropped to be about 8-10mm long. This jig (and transistor leg length) is realistic for something like the AF117 but not really good enough for the AF178. I think I need to test it as per the datasheet with the legs trimmed to 3mm. This would allow comparison with the datasheet measurements up to 400MHz for the AF178.
I also want to measure the noise figure of the 70MHz amplifier that uses the AF117. Maybe also design a (stable) reference 10.7MHz amplifier matched to 50R ports that anyone can build and use as a yardstick. I've got a decent Agilent noise source here and I use an ATE based setup with a netbook and my HP8566B analyser to measure NF using the Y factor method. I get repeatable results with this method although it isn't as good as using a modern/formal NF test setup. Quote:
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 23rd Mar 2018 at 4:16 pm. |
|
23rd Mar 2018, 5:59 pm | #58 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
I did some initial testing of the AF178 and the results are a bit confusing. I crudely measured the current gain at DC to be about 250 and this seems high for a device like this?
The Vbe voltage is typically 0.3V so this is a Ge part and it looks old and genuine. When tested at RF, beta looks to be about 225 at lowish frequencies and this kind of agrees with my DC tests. With my s2p data I measured Ft to be just under 145MHz at 12V 1mA and this is lower than the datasheet figure of 180MHz. The datasheet hints that the gain can be >10dB (13dB typical) at 200MHz. My GMAX result is just under 12dB at 200MHz but if I put in a -5nH inductor in the emitter in Genesys (to make up for the long legs in my test jig) the number rises to about 12.2dB at 200MHz. So this seems about right. So it probably is a genuine AF178 but the Ft seems a tad low and the current gain at DC seems higher than expected?
__________________
Regards, Jeremy G0HZU |
23rd Mar 2018, 10:42 pm | #59 | |
No Longer a Member
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
|
Re: Testing unknown transistors for max frequency response
Quote:
Its sounds as though your one is genuine. As I recall when I tested a batch of AF178's they did have a higher range current gain, but not that high. It would be worth checking one from a different source/batch. Using the rough estimate method from my jig in the common emitter mode the AF178 appeared to have about the right high frequency response(using an AF185 as a reference transistor for the test) at least for the one I tested, but your jig/test is much better. Are you able to estimate the feedback capacitance ? would help confirm its a good genuine AF178? Last edited by Argus25; 23rd Mar 2018 at 10:52 pm. |
|
24th Mar 2018, 4:27 pm | #60 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Testing unknown transistors for max frequency response
If I try and measure it directly via the 12V 1mA s2p model for the AF178 I get about 0.8 - 0.9pF. This seems a bit low but I'm not sure how valid this measurement method is. But I did this in Genesys rather than try and measure it directly.
See below. The data is noisy because I took s2p data at a -35dBm drive level from the VNA.
__________________
Regards, Jeremy G0HZU |