D1 + TDA1541A: can't achieve "0" VDC input
With respect to coupling the D1 with a TDA1541A, Promitheus suggested:
I can't adjust the input to lower than ~25mV. The two variable Bourns pots don't seem to react, even to many turns in either direction. Both sides act symmetrically (i.e., ~25mV on both inputs).
I have noticed some sort of thermal effect. The input DC level falls on its own, but at a rate of about 0.1mV/min.
Bottom line question: for use with a TDA1541A S2 (double crown) in a Philips CDB650, is ~20mV "safe" -- both in terms of chip stability and ultimate audio fidelity?
(Nope: I'm not connecting the D1 to the TDA1541A or CD player to a pre-amp until enough DIYers say it is!).
With respect to coupling the D1 with a TDA1541A, Promitheus suggested:
I can't adjust the input to lower than ~25mV. The two variable Bourns pots don't seem to react, even to many turns in either direction. Both sides act symmetrically (i.e., ~25mV on both inputs).
I have noticed some sort of thermal effect. The input DC level falls on its own, but at a rate of about 0.1mV/min.
Bottom line question: for use with a TDA1541A S2 (double crown) in a Philips CDB650, is ~20mV "safe" -- both in terms of chip stability and ultimate audio fidelity?
(Nope: I'm not connecting the D1 to the TDA1541A or CD player to a pre-amp until enough DIYers say it is!).
These currents are about right for both channels and if the rail voltages are okay, the problem seems to be at the regulator and it's overheating.
The lack of adjustment is another thing. The Source of T1 is to be at 0Volt and is achieved by adjusting the gate voltage via the pot & resistor divider between the +ve rail and gnd - either the pot is faulty (common enough) or the divider is incorrect.
This assumes that the Fets are all okay and if the 2 "arms" of the cct are carrying about 9mA each, any erroneous voltages are fairly clear to see (ie about 13 volt drop on the 1.5kR resistors).
My 2 cents ...
The lack of adjustment is another thing. The Source of T1 is to be at 0Volt and is achieved by adjusting the gate voltage via the pot & resistor divider between the +ve rail and gnd - either the pot is faulty (common enough) or the divider is incorrect.
This assumes that the Fets are all okay and if the 2 "arms" of the cct are carrying about 9mA each, any erroneous voltages are fairly clear to see (ie about 13 volt drop on the 1.5kR resistors).
My 2 cents ...
Couple of other things I should've noted: I had a SNAFU -- short between + and GND ... the 317 beame very hot, and there was smoke around the damper Rs the GND AC (tranf in). Dunno exactly wherethe short occureed, tho'. The PS board for the IV is in an external box, DIN-cable connected to a CDP. I think the short happened "around" the CDP.jameshillj said:
I did replace he 317 and the damper R. After this, I tested the PS: both the + and - sides are well regulated with 1.0mV ripple.
I also ckd out the IV board to some degree -- the input 0VDC adjustment was fine. Sound quality remains "good".
Yeah, this one's got me. But I measured the current draw at the output of the 317. 37 mA it is. I also use the finger-heat test -- keep a finger on the heatsink of the reg while opening and closing various legs. Cutting the input to the IV board cools the reg back to room temp.woodturner-fran said:Wouldn't have thought that 37mA would cause the LM317 to heat
Can others using this IV+PS comment? Thx!
Group BuyAs your currents are okay and rail voltages are also okay, it seems rhat the cct is functioning as it should, apart from the lack of offset.
Last point first - the first suspect is fault pots, or the resistor divider - replace 1 of the pots and the 2 resistors, and check the gate voltage of the T1
317 - simply try a new one!
As you've had a problem with inout supply, I'd also check each of the diodes - actually, if you're using basic ones, suggest replace them with really good ones like BYV17, BAV10, BAT48, SF14, etc.
I'm playing about with Salas' shunt reg at present and will later use this here - maybe get a board from the "B1 + Salas Reg" Group Buy Group Buy - it's perfect for this - expect a HUGE improvement in the sound than basic 317/337 supply.
Last point first - the first suspect is fault pots, or the resistor divider - replace 1 of the pots and the 2 resistors, and check the gate voltage of the T1
317 - simply try a new one!
As you've had a problem with inout supply, I'd also check each of the diodes - actually, if you're using basic ones, suggest replace them with really good ones like BYV17, BAV10, BAT48, SF14, etc.
I'm playing about with Salas' shunt reg at present and will later use this here - maybe get a board from the "B1 + Salas Reg" Group Buy Group Buy - it's perfect for this - expect a HUGE improvement in the sound than basic 317/337 supply.
Haven't done these, so I'll put them on my list. But I'm skeptical: wouldn't problems here affect SQ or the input 0VDC adjustment -- both of which were/are fine.jameshillj said:
As noted prev., I did that. In fact, I did that 2x. And I've used diff. brands: TI, ST.
I think they're okay -- I use Fairchild FFP04S60S 4A, 600V STEALTH II. Wouldn't bad diode(s) affect performance and/or the 337 as well? The protection diodes are common type. Haven't ck'd them.
Got a link on that -- one that shows the schematic?
On shunt-based regs, you may want to look into these.
Back to main issue...
It could be that all is "normal" or has been disfunctional since original build. As I said, I never really gave the regs attention 'till the short-ckt SNAFU.
One other thing: Raw DC after the rectifier/smoothing caps is ~41 V. And the 317 is, of course, "programmed" to output 30V. So that overhead -- in addition to the normal 37 mA current draw for the + side -- may be the reason for a hot-running 317. Dunno why the 337, which has the same voltage overhead, but only 1/2 the current demand (17 mA, see earlier post), doesn't warm up a little up?
Can I convince any fellow Promitheus IV users put a finger on their operating 317 / 337?
Is any warm / hot?Sorry, I thought you had trouble adjusting your dc offset below 25mV and it didn't matter what you did with turning the pots ...
Incidently, the max dc offset for the o/p pins on the 1541A is supposed to be <25mV - not sure where the cutoff point regarding performance is.
Neither my 317 or 337 get at all hot although they're only dropping about 6volts as the basic supply includes a simple Cmultiplier - mind you, yours aren't exactly doing much more 41v -> 30v = 11 volt drop @ 37mA is less than 1/2 watt - nothing to get hot about at all really - don't think the problem is here - not much help, am I!
Yeah, the guy Waagbo developed those regs as a variation and in consultation with Borbelly - dunno where the talk appeared about the simple zener sounding better, but in some circumstances, the regulated supplies seem to choke the sound a bit and some people just like a supply that's got the ripple cap sound - horses for courses, and all that.
I do find that with good diodes, Ripple caps, R-C-R-C system, a Cmultiplier plus some sec winding tuning and well sorted snubbers can result in a really good sound - not the cheapest way to go, tho.
The Salas shunts generally have rather good sound mainly because they were designed to be low noise and there is as much, or more, current going thru the shunt to gnd as there is to the load and this means you can't skimp on the trannie, caps, etc - classic areas where many supplies, shunt or not, are weak.
I haven't yet worked out how to add links here - last night I finally managed to add some "bold" letters!!
The thread is "Simplistic Salas low voltage shunt regulator" - usually has comments every day ...
The group buy is in that section under "DC Coupled B1 buffer with shunt PSU"
I'll see if I can find the 30 volt Salas shunt and PM it to you.
Curiously, I recently heard one of those cheap Blue Ray playback things - picture was REALLY good, but the sound had that same harsh edge to it - long live the good ol' D1 I/V, particularly with a well done non OS player!!
Incidently, the max dc offset for the o/p pins on the 1541A is supposed to be <25mV - not sure where the cutoff point regarding performance is.
Neither my 317 or 337 get at all hot although they're only dropping about 6volts as the basic supply includes a simple Cmultiplier - mind you, yours aren't exactly doing much more 41v -> 30v = 11 volt drop @ 37mA is less than 1/2 watt - nothing to get hot about at all really - don't think the problem is here - not much help, am I!
Yeah, the guy Waagbo developed those regs as a variation and in consultation with Borbelly - dunno where the talk appeared about the simple zener sounding better, but in some circumstances, the regulated supplies seem to choke the sound a bit and some people just like a supply that's got the ripple cap sound - horses for courses, and all that.
I do find that with good diodes, Ripple caps, R-C-R-C system, a Cmultiplier plus some sec winding tuning and well sorted snubbers can result in a really good sound - not the cheapest way to go, tho.
The Salas shunts generally have rather good sound mainly because they were designed to be low noise and there is as much, or more, current going thru the shunt to gnd as there is to the load and this means you can't skimp on the trannie, caps, etc - classic areas where many supplies, shunt or not, are weak.
I haven't yet worked out how to add links here - last night I finally managed to add some "bold" letters!!
The thread is "Simplistic Salas low voltage shunt regulator" - usually has comments every day ...
The group buy is in that section under "DC Coupled B1 buffer with shunt PSU"
I'll see if I can find the 30 volt Salas shunt and PM it to you.
Curiously, I recently heard one of those cheap Blue Ray playback things - picture was REALLY good, but the sound had that same harsh edge to it - long live the good ol' D1 I/V, particularly with a well done non OS player!!
I just recalled I had a similar heat-related issue with another 317-based reg; and that was almost entirely due to extra overhead voltage.
Although I've got some NOS (diy) gear around, I'm generally not a fan of NOS sonics. Your comment does raise a valid question: Do D1 I/V's, and in particular the Prometheus design, sound best with (or are optimized for) NOS?
H.M, sorry can't be of more use with problem.
I think that original was designed for the Pass D1 dac that used the PCM63 chips with 4 X o/sampling and the promitheus version (the one we're using) was also intended for the standard 4 x O/Sampling.
Many people have used it as NOS with good results - me, too.
I did change some of the components like the recommended 2700pF silver, the wima 10nF, etc to mostly styrene, the electros to Silmics, silver wire, etc. as I could optomise the sound of this little cct quite a lot without loosing the ability to bury big amounts of recorded "fatigue"and not loose top end detail.
Curiously, John from EC Design has developed some remarkably simple ccts for the o/p of his new version of the 1541A Nos dac - one has a lot in common with the D1, altho transistors, not fets.
I actually changed the "standard" supply to a "Minnesotta" shunt psu kit, and the improvement is remarkable. I expect the Salas shunt to significantly improve on this, again.
I see you're into headamps - try the EUVL "Taylored DAO Source Follower" - I'm trying the Salas shunt out on this one and the sound into my AKG k701s, Beyer 880s and old Senn 424s is quite extraordinary. With a valve buffer, it's complete!
I think that original was designed for the Pass D1 dac that used the PCM63 chips with 4 X o/sampling and the promitheus version (the one we're using) was also intended for the standard 4 x O/Sampling.
Many people have used it as NOS with good results - me, too.
I did change some of the components like the recommended 2700pF silver, the wima 10nF, etc to mostly styrene, the electros to Silmics, silver wire, etc. as I could optomise the sound of this little cct quite a lot without loosing the ability to bury big amounts of recorded "fatigue"and not loose top end detail.
Curiously, John from EC Design has developed some remarkably simple ccts for the o/p of his new version of the 1541A Nos dac - one has a lot in common with the D1, altho transistors, not fets.
I actually changed the "standard" supply to a "Minnesotta" shunt psu kit, and the improvement is remarkable. I expect the Salas shunt to significantly improve on this, again.
I see you're into headamps - try the EUVL "Taylored DAO Source Follower" - I'm trying the Salas shunt out on this one and the sound into my AKG k701s, Beyer 880s and old Senn 424s is quite extraordinary. With a valve buffer, it's complete!
As far as PSUs for this IV...I'm sure there are far better choices than the default supply that is offered. E.g., Johan's from the link I posted earlier. One may also want to check out Martin's Acoustica site or the PFM forum for the TeddyReg.jameshillj said:
As far as headphone amps: got quite a few of them around. I modded a Meier Corda II -- that one is nice with AC PSU. Built the Tangent PPA2 (the best I've heard). Also modded Go-Vibe, and SuperMacro (fixup.net).
The D1 I/V was designed for balanced operation and when run in that configuration the common mode rejection of the following input stage will remove a significant amount of PSU noise. Depending on the CMRR of the following input stage that could result in somewhere between 30 and 80dB attenuation of common mode noise. This is part of the reason the original D1 I/V could "get away" with using a single pair of 7815/7915 regs to power both channels. I suspect Wayne Colburn's implementation of the 78/7915's actually results in lower noise levels than the 317/337's.
In single end configuration you are relying entirely on the Regs and PSU to eliminate noise.
In single end configuration you are relying entirely on the Regs and PSU to eliminate noise.
Yes, it's still a bit of a surprise but lots of things in power amps just aren't anywhere good enough in headamps.
I never did get around to trying Johan cct - the one with the pre reg LM 317s, but had a good look at the one called UGS Rev3 that seemed interesting with a "quieter" voltage reference - was thinking, could possibly scale it up for the F3 (42V, 1,7A) - also, maybe a power version of Teddy Pardu's Cmultiplier.
Thanks for the info & links - must find out how that "link thing" works - very useful.
Interesting thing is these supply variations all can be readily checked out on the good ol' D1 o/p stage with consistent results.
I wonder if the guys that use IC o/p stages get anything like these results?
I never did get around to trying Johan cct - the one with the pre reg LM 317s, but had a good look at the one called UGS Rev3 that seemed interesting with a "quieter" voltage reference - was thinking, could possibly scale it up for the F3 (42V, 1,7A) - also, maybe a power version of Teddy Pardu's Cmultiplier.
Thanks for the info & links - must find out how that "link thing" works - very useful.
Interesting thing is these supply variations all can be readily checked out on the good ol' D1 o/p stage with consistent results.
I wonder if the guys that use IC o/p stages get anything like these results?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.