Did you reverse engineer a schematic (sketch a schematic from looking at your board and then only after that compare it) for quality control?
Did you check for fake devices, perhaps by swapping in some other suitable devices to see if the large offset problem goes away?
If you visually compare the new board with your working board, what are the differences?
Did you check for fake devices, perhaps by swapping in some other suitable devices to see if the large offset problem goes away?
If you visually compare the new board with your working board, what are the differences?
After posting that, I immediately checked that track again and I saw that my "pull up" method with soldering something didn't get rid of that track properly. So, I removed it by scratching with a tiny screwdriver as what I have to do in first. But, when incidents going to be happen, they are coming faster when you are moving faster. It was totally out of my mind that I removed Q12 and Q13 on pcb. So, guess what happened 
..
I have concentrated myself to build a fresh one already. It is not a big of deal for me anyway. I hope that I will last it this time.
..I have concentrated myself to build a fresh one already. It is not a big of deal for me anyway. I hope that I will last it this time.
Perhaps the next one should use NXP or Fairchild BC560C, BC550C for input?
With the first amp knocked out at Q12 or Q13 and the CFP amps having trouble with BC557, BC547, BC546, I'd have to say that it is probable that you have some counterfeit (or corroded) transistors causing some of these troubles. I also managed to acquire some counterfeit devices since they are so difficult to avoid, especially when shopping for new old stock parts.
P.S.
As for good news, Fairchild, Hi-Sincerity, and Toshiba have PNP drivers with about 26pF that "may" be almost as nice as a Philips BD140.
NXP and Vishay have BAT86 in current production.
ST has 2N3109 (12pF) in current production which is good if your rails are less than 40v; and, Central Semiconductor has 2N3501 (8pF) for $5.50 each and probably workable for high voltage rails.
Fairchild has BC550C, BC560C in current production and also Fairchild has KSC1845 suited for Q12, Q13, in current production.
International shipping from Mouser is $30; however, this is economical since they'll ship a lot of authentic devices in one box.
With the first amp knocked out at Q12 or Q13 and the CFP amps having trouble with BC557, BC547, BC546, I'd have to say that it is probable that you have some counterfeit (or corroded) transistors causing some of these troubles. I also managed to acquire some counterfeit devices since they are so difficult to avoid, especially when shopping for new old stock parts.
P.S.
As for good news, Fairchild, Hi-Sincerity, and Toshiba have PNP drivers with about 26pF that "may" be almost as nice as a Philips BD140.
NXP and Vishay have BAT86 in current production.
ST has 2N3109 (12pF) in current production which is good if your rails are less than 40v; and, Central Semiconductor has 2N3501 (8pF) for $5.50 each and probably workable for high voltage rails.
Fairchild has BC550C, BC560C in current production and also Fairchild has KSC1845 suited for Q12, Q13, in current production.
International shipping from Mouser is $30; however, this is economical since they'll ship a lot of authentic devices in one box.
As you might know, I have also working builds without any trouble since first run. They have same parts populated. There is always a possibility related with selection of parts but this time it was totally related with an accident.
Btw, sometimes accidents causing to good things to be happen. I applied Elvee's offset trimmer on pcb as an optional feature within a very DIY attitude. I'm planning to test it with this build also.
Btw, sometimes accidents causing to good things to be happen. I applied Elvee's offset trimmer on pcb as an optional feature within a very DIY attitude. I'm planning to test it with this build also.
Last edited:
Do you mean you have too much or too little to choose from?I'm struggling to find C4-27pF and C12-820pF. Any suggestions or alternatives. I work for RS Components so have a few to choose from
Anyway, these capacitors are not directly critical, they mainly need to be in the right range, and not of a too bad quality.
Since C4 has such a low value, it will be difficult to be wrong: no risk to find an electrolytic with this kind of value. Any ceramic with a defined tempco, silver mica, glass, styroflex, PP, etc is suitable (and overkill).
For C12, the choice is wider. In principle, "noble" dielectrics of the above-mentioned variety are preferable, but the actual impact on performances is very small, and to keep in with the general philosophy of the Circlophone concept, I took great care to use only X7R grade ceramics on all of my prototypes.
You are of course perfectly free to choose other alternatives, including polycarbonate, polysulfone.....
In fact, you just have to avoid high-k, non-linear dielectrics, which is not difficult for this kind of value.
Any improvement can only bring better results than advertised.
I'm struggling to find C4-27pF and C12-820pF. Any suggestions or alternatives. I work for RS Components so have a few to choose from
C4, Center tolerance at 26pF, so 30pF 27pF, 25pF, 22pF all work.
C12, Most frequently used values by constructors include: 330pF, 470pF, and more. A wide range of component values will work at C12. Apparently, 470pF is center tolerance.
Result:
C4 = 25pF
C12 = 470pF
Last edited:
I tested the idea of weeding out fake transistors by pushing them to the limit.
Basically, the circuit is a CCS connected to a lab supply adjusted to limit at ~250mA.
The transistor itself is mounted on a ~infinite heatsink: a 5mm thick aluminium slab, thermal paste and and tight clamping using proper hardware.
The operating point is choosen to be well inside the SOA of the real thing, here 160mA and 62V.
With this setup, it took less than 10 minutes for the fake to throw in the towel and go short.
The method is simple and accessible to those having limited test equipment.
Basically, the circuit is a CCS connected to a lab supply adjusted to limit at ~250mA.
The transistor itself is mounted on a ~infinite heatsink: a 5mm thick aluminium slab, thermal paste and and tight clamping using proper hardware.
The operating point is choosen to be well inside the SOA of the real thing, here 160mA and 62V.
With this setup, it took less than 10 minutes for the fake to throw in the towel and go short.
The method is simple and accessible to those having limited test equipment.
Attachments
Elvee, I have an extreme curiosity about KSA1220AYS from Mouser.com. Is it probably suitable? In their lineup, it was the closest thing I could find to a Philips BD140. If that sub is is a good idea (if it is better than generic/Motorola BD140), then it would make an entirely Mouser parts list possible in just one box.
It looks ideal: it is equal or superior to the BD140 in all respects.
Simply reverse the supply and auxiliary transistor's polarities.
For the 2SB649, voltage and current are identical, but each type has to be tested in accordance with its SOA, obviously.
I'm still working on documentation pieces. The following is in draft condition.
List for RS and Mouser
D4, D5: MBR735
D7: BAT86
Q1, Q2, Q7: BC556B
Q4, Q3: BC560C (reasonably matched would be good)
Q5, Q6: ST 2N3109 (40+40vdc rails abs max, usable at 37+37vdc)
Q11 Q12: At RS, 2SC2240, At Mouser, KSC1845
Q9, Q11: At RS, 2SA968, At Mouser, 2SA1930 or KSA1220AYS
Q8, Q10: BD911 or MJ15015 or MJL21194, or MJ21194 or MJ15022
Q5, Q6 for higher voltage rails (alternate vendors):
Current: HSD669a from AudioWind (not yet tested), 2SC3421 (not yet tested)
NOS: 2N3498, 2N3499, 2N3500, 2N3501, BF819
Limited supply, 89 of Central Semi 2N3501 is in stock at Mouser for $5.20 each. Limited supply of 2N3501 is in stock at Affiliated Electronics.
List for RS and Mouser
D4, D5: MBR735
D7: BAT86
Q1, Q2, Q7: BC556B
Q4, Q3: BC560C (reasonably matched would be good)
Q5, Q6: ST 2N3109 (40+40vdc rails abs max, usable at 37+37vdc)
Q11 Q12: At RS, 2SC2240, At Mouser, KSC1845
Q9, Q11: At RS, 2SA968, At Mouser, 2SA1930 or KSA1220AYS
Q8, Q10: BD911 or MJ15015 or MJL21194, or MJ21194 or MJ15022
Q5, Q6 for higher voltage rails (alternate vendors):
Current: HSD669a from AudioWind (not yet tested), 2SC3421 (not yet tested)
NOS: 2N3498, 2N3499, 2N3500, 2N3501, BF819
Limited supply, 89 of Central Semi 2N3501 is in stock at Mouser for $5.20 each. Limited supply of 2N3501 is in stock at Affiliated Electronics.
You mean 2N3019... If you could find some 2n5858's, I have to say that it is a very successful candidate. You could give a try if you can find some.
Last edited:
Thanks for catching the transpose error so quickly. Q5, Q6, Yes, I meant 2N3019--Both Mouser and RS have ST's version, 12pF.
Littlediode carries Central Semiconductor 2N5858 which is $6.38 each parts plus approximately $4 shipping (depending on location). With these located in London, shipping there may be much less. Those devices are very fast.
Have we ever tried the FET alternative?
That's really expensive! What is wrong with these good transistors those aren't in production line in these days? (or decades?)
For myself, my priority is bjt. If I couldn't find anything in place of Q5 and Q6 (which is not a easy case), then mosfets may come to next to try. Although, it will be interesting to see anyone used mosfets in place of Q5/Q6 successfully.
In sim, it works absolutely transparently, only the values of R4 and R13 need to be adapted to suit the different threshold voltage, but other than that, it looks like a simple plug-in replacement.
Any small MOS of sufficient voltage should do.
I'll give it a try one of these days.
Attachments
"Eco-friendly Halogen-free package"
If there was enough space to place some Halogen in that package..