One characteristic I place some high importance on - more than most others - is how flat the hfe curve is with varying Ic. If you have a pair that are as flat and the Vbe is .72 v AND the hfe match and the temp coeff are complimentary, then they are better.
The power supply IS important. 3-terminal regs are the minimum performance. And, in this app dont need to be pre-loaded to get the regulation to work best. But you should strive for better. Shunt reg is my preferred method. To get up and runnning, the 3-term gets you there quickly.
-Richard Marsh
The power supply IS important. 3-terminal regs are the minimum performance. And, in this app dont need to be pre-loaded to get the regulation to work best. But you should strive for better. Shunt reg is my preferred method. To get up and runnning, the 3-term gets you there quickly.
-Richard Marsh
Last edited:
Hi homemodder, thanks for your comments.
I agree with you about the MJE172/182 but unfortunately I have not have any of them. Regarding the 2sa1930/2sc5171, they are quite similar to the ones I mentioned in my previous post, 2sa1306/2sc3298 so they have similar problems: different body, different pin-out and quite dissimilar electrical characteristics. But there are ways to use them: for instance I can mount them on the bottom of the PCB, but I still have to conform the pins as they measure differently.
In conclusion, I think my best option is the BD139/140.
Javier
Pulllleeeaaase, try the LT1963A/LT3015 -- this is the best sounding pair of OEM regulators, hands down.
dual secondaries allow using the same Vreg part type for plus and minus - you can use the better reg (lower noise, better ripple rejection, load step recovery...) for both, in duplicate reg circuits
cuts the number of different part types, you only have to "optimize", characterize one circuit design instead of 2
cuts the number of different part types, you only have to "optimize", characterize one circuit design instead of 2
Last edited:
Dont bother with the 317/337 at all. Use the Linear Technology for 3-terminal regulators for thier convenience/simplicity of built.... and better performance.
I bought about 20 each of the Hitachi 2SD669A/2SB649A devices. Very consistant... They both test .72v and Hfe within 5% of compliment right out of the box. They look good down to 30 Ohm loads. I havent tried them in amp but should be a drop-in replacement.
[the MJE-xxx can go down to lower Z loads than 30 Ohms, if that is important to you.] -RNM
[the MJE-xxx can go down to lower Z loads than 30 Ohms, if that is important to you.] -RNM
Last edited:
Could you tell us where did you get them? Just because as I've said before the ones I got were not very good complementaries (Vbe for PNP is 0.74v and for NPN 0.80v).
Javier
I dont recall..... project is a year old for me. But I got them on eBAy and I believe you can get them matched from 'tubeshunter' out of hkaudioparts. he sells on eBay.
I just looked on eBay for the matched pair and found a set of 4 matched ones by same person for sale there right now.
I just looked on eBay for the matched pair and found a set of 4 matched ones by same person for sale there right now.
Last edited:
The hitachi pair has been discontinued for about 10 years now, the chance that anyone is selling real hitachi is virtually nil. The ones by tubehunter are not hitachis but imitations. You might think they good because he matches them and it helps to sell but those are at best manufactured second source (trying to keep close to spec) or more realistic, as they use the hitachi logo to make believe they are originals, cheap makeovers probably manufactured in Hong Kong by SECOS (These guys put any logo you may wish for on components). In most cases these components dont perform to Hitachi spec although there are two manufacturers which produce pretty good copies, one company is from UK (Transys) and the other a company in Taiwan (Unisonic). The rest, well, youll have to put them on a curve tracer and see. Some time back I tried some pairs from SVNTC (India) with mixed results, some up there to spec some completly off.
PS The 2sb631 pair is very good too, sadly discontinued.
PS The 2sb631 pair is very good too, sadly discontinued.
The BDs is what I would use if they were from manufacturer like ONsemi, Fairchild, STi or Phillips. These are excellent components for this type of circuit.
If you want to get the hitach pair buy them from Profusion Plc in the UK. They sell them for 14 or 15 cents each so you can buy many to find matches. Note these are second sourced, see my previous post regarding these components. Dont get fooled by imitations on Ebay where you pay 90c or a dollar for thinking they are genuine, the seller is making a killing.
If you want to see what a genuine Hitachi looks like drop me a pm.
Why speculate on thier suitability -- Ill put them on a curve tracer and evaluate. Using the simple Peak Atlas tester, they had better Hfe than the BD types and the matching was very tight -- a savings in time, if not money.
The matching of compliments and topology matters to get lowest thd. Having said that - For no dc output drift.... Dont forget this aspect if you want to try some other devices.... temp coefficients.
The matching of compliments and topology matters to get lowest thd. Having said that - For no dc output drift.... Dont forget this aspect if you want to try some other devices.... temp coefficients.
Im not speculating, those parts by tubeshunter are not genuine hitachi although they will work, so will another 50 different pairs of transitors accross 30 manufacturers. Obviously its possible to find matching in any transistors be they even from fake manufacturer, it only takes time. What Im saying is that they might not follow spec according to hitachi spec sheets. I dont know where you got those BDs from, again from Ebay ?? I have many BDs genuine sti and philips and they have higher HFE (greater than 140) than all mje172/182 I have. It all depends on the batch and the company reputation that manufactured. Profusion s second source hitachis all have a min hfe of 160.
@jackinnj
Looking at the datasheet for the LT1963, I'm leaning toward a tantalum cap (SMT) with low ESR, say about 22uF. Do you have a better recommendation? I'm ok with SMT as long as package isn't about the size of this period at the end of my sentence.
Javier, if this is the cap after the reg, you normally don't want it to be super-low ESR. The ESR part is required for loop damping to enhance stability.
Very high quality film caps are a no-no here.
I have a prototype 400VDC reg where I actually had to insert a 1 ohm R in series with the output cap to quiet it down - my electrolytic was simply too good
jan