See an example. http://www.diyaudio.com/forums/pass...ymmetrical-psu-b1-buffer-181.html#post2247197 The inner ones are to be the slightly less in IDSS per pair, and the pairs should be around same ballpark not to have much difference in dissipation and to behave similar dynamically.
OK, I'm coming back to the well one more time to be sure I understand this. In trying to select parts it is difficult to find transistors which list hFE on the order of 10; for example, the 2N3904 data sheet (http://www.fairchildsemi.com/ds/MM/MMBT3904.pdf) lists hFE's of 30-300. However, those are all at Vce=1.0V and varying collector currents. Then it shows collector-emitter saturation voltage of 0.2V, with Ib=1.0mA and Ic=10mA, which to me implies an hFE of 10. Further, it lists the base-emitter saturation voltage of 0.65-0.85V. Can I take it that these "saturation voltages", as opposed to "breakdown voltages", are what I am interested in, and that hFE is not some fixed gain ratio but a more fluid quantity that varies a lot with other operating conditions? Would the 3904 be a good choice? Would a Darlington with higher current gain be better? The 3904 seems better to me than the 2N2222 because of the lower base current, and the lower base-emitter saturation voltage.
I really do appreciate your pointers, and hope I'm not hijacking this thread with what is really a side-bar.
Thanks.
Hi,
yes and no.
Use any cheap To92 that has sufficient Ic max. Speed, Vce0 and gain are unimportant in this slow switching application.
Using an excess of base current one pushes the switching transistor into saturation, where Vbe is higher than the usual 650mVbe +-50mV and the Vce is well below 200mV. At an effective 1:10 ratio of base:collector current the Vce can usually be in the region of 1mV to 50mV.
But the transistor has no internal mechanism to protect the base from damaging base current values. You must have a base current limiter. This is usually just a resistor that limits the worst case base current to ~10% of the collector current.
yes and no.
Use any cheap To92 that has sufficient Ic max. Speed, Vce0 and gain are unimportant in this slow switching application.
Using an excess of base current one pushes the switching transistor into saturation, where Vbe is higher than the usual 650mVbe +-50mV and the Vce is well below 200mV. At an effective 1:10 ratio of base:collector current the Vce can usually be in the region of 1mV to 50mV.
But the transistor has no internal mechanism to protect the base from damaging base current values. You must have a base current limiter. This is usually just a resistor that limits the worst case base current to ~10% of the collector current.
I’ve been looking for your source to find out more info, just to confirm the 20R/2W x 2 units replaces the 4 x 68R/0.5W is that correct?
There are two CCS in the Mez, the +-ve supply and the -ve supply.
Each supply has two locations for resistors. That is 4 resistors to each Mez PCB.
I use 600mW resistors, my stock of power resistors is tiny.
I find that 30r with ~1.8V across it dissipates ~110mW. this will run fairly cool.
I also use the bottom side of the PCB to add extra parallel resistors, if I need more current.
Now you could use up to 8off 30r resistors on 1 PCB.
Each 30r gives about 60mA of CCS current. This needs to be checked at time of build.
If you fit 3off 30r you will get ~3times the current.
If you fit 7off 30r you will get ~7times the current.
Using this methodology I don't need to buy any new resistors, they all come from my standard E24 stock.
Each supply has two locations for resistors. That is 4 resistors to each Mez PCB.
I use 600mW resistors, my stock of power resistors is tiny.
I find that 30r with ~1.8V across it dissipates ~110mW. this will run fairly cool.
I also use the bottom side of the PCB to add extra parallel resistors, if I need more current.
Now you could use up to 8off 30r resistors on 1 PCB.
Each 30r gives about 60mA of CCS current. This needs to be checked at time of build.
If you fit 3off 30r you will get ~3times the current.
If you fit 7off 30r you will get ~7times the current.
Using this methodology I don't need to buy any new resistors, they all come from my standard E24 stock.
I just finished building a Mezmerize dcb1 and powered it up for the first time. The power LED lights up as well as both clusters of 5 LEDs, but the two clusters of 3 LEDs don't appear to light up at all. I hear one of the omron relays click after a few seconds. The LEDs came from Ed Lafontaine's group buy. I built it per the BOM except with a 12v transformer.
I haven't done any measurements or tests yet, I'll check the DC offset and whatnot once I remember where I put my DMM.
I haven't done any measurements or tests yet, I'll check the DC offset and whatnot once I remember where I put my DMM.
Whoops, I didn't mean to even temporarily impugn the quality of your LEDs, Ed. Everything you sent works 100% when you don't plug it in backwards.
I plan to eventually use a Lightspeed Attenuator with this puppy, and I'd like to power it off the 12v regulator on the DCB1. Does anybody forsee any issues with that? I have no idea how much current that circuit draws.
I plan to eventually use a Lightspeed Attenuator with this puppy, and I'd like to power it off the 12v regulator on the DCB1. Does anybody forsee any issues with that? I have no idea how much current that circuit draws.
Thanks to AndrewT I will shortly be able to make up either two pairs of matched 7.8mA SK170's, or combine one pair of matched 8.4mA with one pair of 8.6mA Sk170's.
Having read and digested the various threads on the B1 and the DCB1 it would appear that having a higher mA rating is better than a lower unit, but also having as closely matched pairs as possible is also better.
My view on this, having read the details, is using the 8.4mA and 8.6mA matched pairs will provide more benefit than using the lower 7.8mA macthed pairs, even though there is 0.2 mA difference between the 8.4mA and 8.6mA pairs. It would appear that a difference of 0.2 mA is easily tolerated and thus the higher current rating is preferred.
Would you agree with my conclusions above?
I also believe that I would need to install them as follows: 8.4/8.4/8.6/8.6 as per this link http://www.diyaudio.com/forums/pass-labs/145201-building-symmetrical-psu-b1-buffer-181.html#post2247197 where the 8.4 pair would be in positions 1 and 2, and the 8.6 pair would be in positions 3 and 4.
Would you agree with that also?
Thanks
Having read and digested the various threads on the B1 and the DCB1 it would appear that having a higher mA rating is better than a lower unit, but also having as closely matched pairs as possible is also better.
My view on this, having read the details, is using the 8.4mA and 8.6mA matched pairs will provide more benefit than using the lower 7.8mA macthed pairs, even though there is 0.2 mA difference between the 8.4mA and 8.6mA pairs. It would appear that a difference of 0.2 mA is easily tolerated and thus the higher current rating is preferred.
Would you agree with my conclusions above?
I also believe that I would need to install them as follows: 8.4/8.4/8.6/8.6 as per this link http://www.diyaudio.com/forums/pass-labs/145201-building-symmetrical-psu-b1-buffer-181.html#post2247197 where the 8.4 pair would be in positions 1 and 2, and the 8.6 pair would be in positions 3 and 4.
Would you agree with that also?
Thanks
Is this just a super modified B1 and has anybody completed a balanced version. I have read a lot, but i can't seem to find the beginning or end. Ordered two hot rod hynotize boarsd just in case i figure it out!
One guy was posting about planning to make a balanced one. Have not seen him posting he finished yet.
I don't see much difference in transconductance for 1mA IDSS steps in the Toshiba chart. So the 7.5-8.5mA region for instance isn't that different. Much more, like beyond 10mA would bring unwanted dissipation. The 8.4/8.4/8.6/8.6 scheme is fine.
That could well be me... I snapped my wrist before xmas, just now back to mostly working, so I've ordered most of the parts. I still need to order the matched FETs, looks like AndrewT might be the man for this. I expect to have the parts in weeks or so, then give me a couple of weeks and I should have it built, unless something else happens!