M3 Semiconductor Analyzer
Hello,
For those not familiar with the M3 semiconductor analyzer, here is a link to the site:
http://www.m3electronix.com/kit.html
Specifically, here is the one I built, documented, and submitted to M3 for posting:
http://www.m3electronix.com/paul.html
I bought one and built it for the great features and low cost. While not perfect and far from an industrial grade, professional test device, it offers the home hobbiest some great features and flexibility in an inexpensive device.
Using the M3 for DIY audio projects should be adequate for some or many of us, but naturally there are those who will scoff at it and will insist on using extreme measures for their work. To each his own. I don't use $2,000 speaker wires either, so what do I know.
Regards,
Paul
Hello,
For those not familiar with the M3 semiconductor analyzer, here is a link to the site:
http://www.m3electronix.com/kit.html
Specifically, here is the one I built, documented, and submitted to M3 for posting:
http://www.m3electronix.com/paul.html
I bought one and built it for the great features and low cost. While not perfect and far from an industrial grade, professional test device, it offers the home hobbiest some great features and flexibility in an inexpensive device.
Using the M3 for DIY audio projects should be adequate for some or many of us, but naturally there are those who will scoff at it and will insist on using extreme measures for their work. To each his own. I don't use $2,000 speaker wires either, so what do I know.
Regards,
Paul
Hfe testing
In manufacturer's high speed autotesters, any test causing a chip temperature rise (Vbe(sat) Vce(sat), Hfe at currents > 100mA) is normally tested with a 300uS (*) pulse of current with a low (< 1%) duty cycle.
Hfe is tested
(a) in an analog servo loop at constant Ic, servoing Ib until Vce is at the required vaue.
Problems here are loop stability varying with device type.
b) using a digital servo binary-splitting Ib and detecting if Vce is > < required value.
Problem here is time, but only 12 iterations (each taking 30mS) will get the result.
Very hard to correleate this to amateur DC measurements better than 10%.
* There is a reason for the 300uS figure based on an early but superb piece of electronic equipment. Not many of us are old enough to remember it ............
In manufacturer's high speed autotesters, any test causing a chip temperature rise (Vbe(sat) Vce(sat), Hfe at currents > 100mA) is normally tested with a 300uS (*) pulse of current with a low (< 1%) duty cycle.
Hfe is tested
(a) in an analog servo loop at constant Ic, servoing Ib until Vce is at the required vaue.
Problems here are loop stability varying with device type.
b) using a digital servo binary-splitting Ib and detecting if Vce is > < required value.
Problem here is time, but only 12 iterations (each taking 30mS) will get the result.
Very hard to correleate this to amateur DC measurements better than 10%.
* There is a reason for the 300uS figure based on an early but superb piece of electronic equipment. Not many of us are old enough to remember it ............
thanks, cliff
for your inside information
and the above info from you means for us ....
As you know these things in such details
you may be the one to answer or suggest something good
for the everyday common diy builder.
What you say this means we could do for our own practical testings?
thank you
regards
lineup
Lineup Audio DIY Lab - specializing in transistor amps
for your inside information
and the above info from you means for us ....
As you know these things in such details
you may be the one to answer or suggest something good
for the everyday common diy builder.
What you say this means we could do for our own practical testings?
thank you
regards
lineup
Lineup Audio DIY Lab - specializing in transistor amps
Hi MikeB,
Yes, I find it sad too to have so many pairs of MJE that can't match. For symasym, I think that a possibibility would be to use the D44H11 AND D45H11.
These are available here from ST and they have Vceo at 80 V, Ft is at 50 MHz for 0,5 A. Only, you will not be able to use a 30V-0-30V, but only a 25V-0-25V transformer in the power supply. This is not an issue for symasym.
The gain should be high enough as the min is at 60 for Ic=2A. The MJE15030/1 have here a lower spec having hFE min being min 40 at 2 A.
Models for simulators are easily found for them. I think these are good devices.
Best regards.
rephil
Yes, I find it sad too to have so many pairs of MJE that can't match. For symasym, I think that a possibibility would be to use the D44H11 AND D45H11.
These are available here from ST and they have Vceo at 80 V, Ft is at 50 MHz for 0,5 A. Only, you will not be able to use a 30V-0-30V, but only a 25V-0-25V transformer in the power supply. This is not an issue for symasym.
The gain should be high enough as the min is at 60 for Ic=2A. The MJE15030/1 have here a lower spec having hFE min being min 40 at 2 A.
Models for simulators are easily found for them. I think these are good devices.
Best regards.
rephil
Hi,
the examples quoted are typical of BJTs.
An increase in Vce0 and some or many of the characteristics are compromised.
The D series being low voltage will probably gain in other respects and you have featured gain and frequency.
BUT,
at this low a voltage, you may find that SOA is also low at medium to high Vce (30Vce to 80Vce). You MUST check the datasheet for SOA at the typical voltages across the transistor in this amplifier.
the examples quoted are typical of BJTs.
An increase in Vce0 and some or many of the characteristics are compromised.
The D series being low voltage will probably gain in other respects and you have featured gain and frequency.
BUT,
at this low a voltage, you may find that SOA is also low at medium to high Vce (30Vce to 80Vce). You MUST check the datasheet for SOA at the typical voltages across the transistor in this amplifier.
Hi AndrewT at the Scottish Borders,
We are speaking in this thread of the transistor hFE observations. So I must not begin to speak of something else.
But I must reply friendly to you. So : the D transistors should be drivers in symasym as the MJE are. With 36V supply, they will never see more than Ic = 200mA. With Vceo at 80V, the D should be fine. All is there. The gain is mentionned only because MikeB has said somewhere else that the gain of the drivers should not be below 100. The fT because it is a nice value.
Please, consider that I am learning, and that nobody should think differently from me.
There are also a lot of things I should like to find here in diyaudio, like a plot of the 3 important curves that are at the centre of the symasym amp : the open loop gain Aol, the inverse of the beta gain curve and the betaxAol curve, this using the Middlebrook technic (Mikeks, Pedja). I have done it here this way at home, and I have still questions without answer about this power amp.
Best regards
rephil
We are speaking in this thread of the transistor hFE observations. So I must not begin to speak of something else.
But I must reply friendly to you. So : the D transistors should be drivers in symasym as the MJE are. With 36V supply, they will never see more than Ic = 200mA. With Vceo at 80V, the D should be fine. All is there. The gain is mentionned only because MikeB has said somewhere else that the gain of the drivers should not be below 100. The fT because it is a nice value.
Please, consider that I am learning, and that nobody should think differently from me.
There are also a lot of things I should like to find here in diyaudio, like a plot of the 3 important curves that are at the centre of the symasym amp : the open loop gain Aol, the inverse of the beta gain curve and the betaxAol curve, this using the Middlebrook technic (Mikeks, Pedja). I have done it here this way at home, and I have still questions without answer about this power amp.
Best regards
rephil
Hi AndrewT at the Scottish Borders,
Correction :
With 36V supply as they should never see more than Ic=200mA (on MikeB load of 8 and 4 Ohm), with the Vceo at 80V the D should be fine.
I did not sleep last night, and now I am out ...
I hope that I did not write something else very wrong.
Best regards
rephil
Correction :
With 36V supply as they should never see more than Ic=200mA (on MikeB load of 8 and 4 Ohm), with the Vceo at 80V the D should be fine.
I did not sleep last night, and now I am out ...
I hope that I did not write something else very wrong.
Best regards
rephil
Hi Rephil
Lots of points to answer! For now - post #16. Yes, I use 4A as a good guide for 50W amps. Lineup already said that you should use a small heatsink to minimise heating effects during measurements.
I would also do this for TO-220 drivers. You won't be able to make a manual measurement fast enough to stop this at high currents otherwise, and 100-200mA are useful currents to be measuring.
For my gain measurements I used an old 5V 20A computer PSU - it is fine for trannys like 2N3771 as well (and I was once sold a pair which looked like 2N3055's in wolfs clothing - gain was less than about 5 at 15A (should have been >=15).
cheers
John
Lots of points to answer! For now - post #16. Yes, I use 4A as a good guide for 50W amps. Lineup already said that you should use a small heatsink to minimise heating effects during measurements.
I would also do this for TO-220 drivers. You won't be able to make a manual measurement fast enough to stop this at high currents otherwise, and 100-200mA are useful currents to be measuring.
For my gain measurements I used an old 5V 20A computer PSU - it is fine for trannys like 2N3771 as well (and I was once sold a pair which looked like 2N3055's in wolfs clothing - gain was less than about 5 at 15A (should have been >=15).
cheers
John
Hi AndrewT,
Yes, I MUST check the datasheet for SOA at the typical voltages across the transistor in this amplifier. It was useful to write it.
I MUST also apologize with you as I did not thank you for your friendly help! I am so sorry for this!
Please don't look at what I buy or I have bought and even forget the way I post here as I am recently coming out of 10 very black years in my life. I have had the death in my home, and until the last Christmas, different surgeons worked many times heavily on me. I am out of my health problems now, but the recovery is slow and painful.
Believe me, the only way I had to survive to all I had to live was to dream to realize my own first 50W power amp, this using my own circuitry in my small lab. So during these years, when they did not cure me, I have planned this dream, buying from time to time what I believed at the moment useful for future use, this with the money I did not spend as I have had no hollydays during all these years ... I was preparing the present days.
I am still dreaming, but now I know that it is possible with your help and with the help of the fine guys that share their knowledge posting here. So I am sure that I will have success some day.
Merry Christmas AndrewT and again many thanks !
Best regards
rephil
Yes, I MUST check the datasheet for SOA at the typical voltages across the transistor in this amplifier. It was useful to write it.
I MUST also apologize with you as I did not thank you for your friendly help! I am so sorry for this!
Please don't look at what I buy or I have bought and even forget the way I post here as I am recently coming out of 10 very black years in my life. I have had the death in my home, and until the last Christmas, different surgeons worked many times heavily on me. I am out of my health problems now, but the recovery is slow and painful.
Believe me, the only way I had to survive to all I had to live was to dream to realize my own first 50W power amp, this using my own circuitry in my small lab. So during these years, when they did not cure me, I have planned this dream, buying from time to time what I believed at the moment useful for future use, this with the money I did not spend as I have had no hollydays during all these years ... I was preparing the present days.
I am still dreaming, but now I know that it is possible with your help and with the help of the fine guys that share their knowledge posting here. So I am sure that I will have success some day.
Merry Christmas AndrewT and again many thanks !
Best regards
rephil
Hi john_ellis,
Thank you for your reply. Nice to see that you were able to use an old 5V 20A computer PSU !
I hope some day to get the necessary skills that will free me of what I consider "to throw money through the window" in buying items that it is possible to do without and still obtaining the same good results.
Merry Christmas to you !
Best regards
rephil
Thank you for your reply. Nice to see that you were able to use an old 5V 20A computer PSU !
I hope some day to get the necessary skills that will free me of what I consider "to throw money through the window" in buying items that it is possible to do without and still obtaining the same good results.
Merry Christmas to you !
Best regards
rephil
I dug out my batch of 15003 devices last night and ran through them.
I have 16 devices and the HFE's ran from 32 to 46 with most being in the mid to high 30's.
This time i payed attention to he Vbe numbers and i noticed that they ran from .52v to .54v with .53v for almost all excpet for the extreme ends of the Hfe scale. But the base current changed for each device.
I will need to experiment with the M3 a bit further to explore the Vbe numbers. and Base current readings.
Zc
I have 16 devices and the HFE's ran from 32 to 46 with most being in the mid to high 30's.
This time i payed attention to he Vbe numbers and i noticed that they ran from .52v to .54v with .53v for almost all excpet for the extreme ends of the Hfe scale. But the base current changed for each device.
I will need to experiment with the M3 a bit further to explore the Vbe numbers. and Base current readings.
Zc
Hi,
when setting up the output and driver transistors, the voltage across the base-emitter plus the voltage across the emitter resistor is effectively constant and equal for all parallel devices.
If the emitter resistor is equal for all, then a variation in Vbe WILL make the higher Vbe devices carry less current and run cooler.
The devices with lower Vbe WILL carry more of the current share and WILL run hotter.
Different temperatures and different currents and different Vbe does not sound like a happy marriage for parallel devices.
It seems to me that the most important characteristic that needs matching is Vbe at Iq and secondly Vbe at some appropriate but higher current. Matching Vbe at Iq ensures that [Vbe + Vre] = constant for all parallel devices AT Iq.
when setting up the output and driver transistors, the voltage across the base-emitter plus the voltage across the emitter resistor is effectively constant and equal for all parallel devices.
If the emitter resistor is equal for all, then a variation in Vbe WILL make the higher Vbe devices carry less current and run cooler.
The devices with lower Vbe WILL carry more of the current share and WILL run hotter.
Different temperatures and different currents and different Vbe does not sound like a happy marriage for parallel devices.
It seems to me that the most important characteristic that needs matching is Vbe at Iq and secondly Vbe at some appropriate but higher current. Matching Vbe at Iq ensures that [Vbe + Vre] = constant for all parallel devices AT Iq.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.