@jean-paul:
Current draw of current production TL072 (older ones may be higher) is shown as between 937.5 and 1125 micro-amps per amplifier, so about 1.8 - 2.2mA per dual package. Datasheet for OPA2134 states quiescent current draw as 4-5mA per amplifer, or between 8-10mA per dual package. If I recall correctly, there are 14 op amps in the APT Holman, but most get only 12 replaced, so let's say the additional current draw is on average, say 2mA/op amp x 12 = 24mA for standard TL072 op amps, vs. 9mA/op amp x 12 = 108mA for OPA2134 - a difference of 84mA. The mains fuse (120V supply) is rated at 125mA.
The APT service manual specifies max current draw of the standard pream at less than 15 watts, or about 0.125A, which suggests they have selected a very small margin for the fuse, and based on my own experience with two different APT transformers (different makes, configurations), they may have also been just up to the task of the standard configuration.
I had one unit with a failing transformer that would provide enough power for the preamp to function only if the relay was taken out of circuit, but when re-connected, the transformer would heat up and the mute circuit would drop out for lack of V and I.
My own experience with the original muting circuit is it works just fine, can be easily repaired, but a high-quality small-signal relay with lower current draw and higher quality contacts and sealed package is a very significant improvement for this preamp. Do it once, and it's good for the remaining life of the preamp, in my opinion. There are wonderful small-signal relays available now. Recently I built a phono-level relay switch/jack box for all of my phono devices to route to up to 3 different preamps, using excellent relays, and a PCB kindly and expertly designed by Sean. These relays are in my view, a great improvement over the standard ones selected by APT. Remember, they had to build these to a price for the larger market of aspiring audiophiles.
Current draw of current production TL072 (older ones may be higher) is shown as between 937.5 and 1125 micro-amps per amplifier, so about 1.8 - 2.2mA per dual package. Datasheet for OPA2134 states quiescent current draw as 4-5mA per amplifer, or between 8-10mA per dual package. If I recall correctly, there are 14 op amps in the APT Holman, but most get only 12 replaced, so let's say the additional current draw is on average, say 2mA/op amp x 12 = 24mA for standard TL072 op amps, vs. 9mA/op amp x 12 = 108mA for OPA2134 - a difference of 84mA. The mains fuse (120V supply) is rated at 125mA.
The APT service manual specifies max current draw of the standard pream at less than 15 watts, or about 0.125A, which suggests they have selected a very small margin for the fuse, and based on my own experience with two different APT transformers (different makes, configurations), they may have also been just up to the task of the standard configuration.
I had one unit with a failing transformer that would provide enough power for the preamp to function only if the relay was taken out of circuit, but when re-connected, the transformer would heat up and the mute circuit would drop out for lack of V and I.
My own experience with the original muting circuit is it works just fine, can be easily repaired, but a high-quality small-signal relay with lower current draw and higher quality contacts and sealed package is a very significant improvement for this preamp. Do it once, and it's good for the remaining life of the preamp, in my opinion. There are wonderful small-signal relays available now. Recently I built a phono-level relay switch/jack box for all of my phono devices to route to up to 3 different preamps, using excellent relays, and a PCB kindly and expertly designed by Sean. These relays are in my view, a great improvement over the standard ones selected by APT. Remember, they had to build these to a price for the larger market of aspiring audiophiles.
Not typically mentioned in the data sheets: it takes less coil current to hold the contacts closed than to pull them closed in the first place (usu. called "pull-in" & "hold-in" voltage/current).
Apt implemented this in the Q19/R90 circuit. In a nutshell, a higher current is provided at first (thru Q19) to firmly close the relay, then R90 kicks in to reduce current to a "hold" value. See para. 4.2 in the manual for full description.
So 820 ohm would allow slightly higher "hold" current than 1k. I can only presume it was done to address a perceived weakness.
A spice graph (from a year or so ago while trouble shooting another users mute circuit):
Green trace is the delay drive voltage to Q17. Red is current thru coil: peaks at 22mA, then drops (as C67 charges) until Q19 turns off, stabilizing at 13mA+ thru R90.
Thanks for the explanation with expected results and the confirmation that suggested less power hungry signal relays do a better job here. Never have had such a preamp (and never will) so acting on experience with similar devices.
BTW the power transformer is chosen with a too low margin (again a mediocre choice) and 84 mA really was the current of a few LEDs then. A 25VA transformer will not be much larger than a 15VA rated one or will I get haunted by the "original" again?
Again I am surprised that brands and their products sometimes have fans that really accept anything. When you have an established reputation you can make many mistakes and the buyers probably will like the products even more. When I started working in audio I already noticed the worst devices with highest failure rates to be most popular. It surprised me and coworkers what reasons consumers gave for buying exactly that device and it rarely were facts, just emotions. Amazing.
Last edited:
@jean-paul:
I understand what you are saying. Both of the APTs I have that needed new transformers got higher VA, better quality units. The first was a high quality, medical-grade shielded toroid, but that required ditching the courtesy outlets on the back (I don't use those anyway) and some bracket fabrication. The second got a traditional EI transformer, but upgraded.
Why do many people like these preamps? Well, keeping in mind when they were made and the lower audiophile price range (higher than mid-fi gear but not as much as most boutique brands), they provided great sound with an excellent suite of features. That was still when full-function preamps were very popular. The performance and sounds was quite good, and can be made even better with some relatively minor upgrades. It easily competes in sound quality, performance and features with many of my other preamps that cost much more back in the day. Some of those have an emotional, faithful following which I don't understand, because I don't enjoy listening to them (e.g. Audible Illusions, and others). I have a lot of nice preamps, and my upgraded APT Holmans give little away in sound quality.
Appearance? Well, not as much.
I understand what you are saying. Both of the APTs I have that needed new transformers got higher VA, better quality units. The first was a high quality, medical-grade shielded toroid, but that required ditching the courtesy outlets on the back (I don't use those anyway) and some bracket fabrication. The second got a traditional EI transformer, but upgraded.
Why do many people like these preamps? Well, keeping in mind when they were made and the lower audiophile price range (higher than mid-fi gear but not as much as most boutique brands), they provided great sound with an excellent suite of features. That was still when full-function preamps were very popular. The performance and sounds was quite good, and can be made even better with some relatively minor upgrades. It easily competes in sound quality, performance and features with many of my other preamps that cost much more back in the day. Some of those have an emotional, faithful following which I don't understand, because I don't enjoy listening to them (e.g. Audible Illusions, and others). I have a lot of nice preamps, and my upgraded APT Holmans give little away in sound quality.
Appearance? Well, not as much.
This is the relay i purchased. Mine fit into the circuit board perfectly.@seanc
I actually ordered one of those last night from Galco, but it seems like the one I bought is actually the PCB version:
https://www.galco.com/ry2v-u-dc24-idec.html
We'll see when it shows up! I also bought another signal relay to try when I was putting together a Mouser order for the components to revert it back to the original muting circuit.
That's a good observation about the new op amps. I installed DIP sockets everywhere so it will be possible to put the old op amps back if needed and/or for comparison.
BTW I think I got the idea for the Samtec ribbon cables from one of your posts, so thanks! It would have been really difficult to troubleshoot this muting problem if the boards were all permanently connected.
These transistors should work better, more current and MW dissipation 2n4403 and 2n4401
What VA did you choose? And are the correct secondaries +/- 25V with a center tap or equivalently dual 25V secondaries tied together with a ground at the tie point? did I say that correctly?
From what I can see, the unregulated +/- 24Vdc is only used for the relay coil and headphone amp op-amp. So could probably feed a lower voltage into the + and - 18V regulators? of course you'd need a relay with lower could voltage and disconnect the headphone amp.
Original transformer voltage is 18-0-18, except that at some point they went to a 15-0-15VAC transformer with no end bells, and clearly not up to the quality standards of the original unit. They had to modify the circuit board for the different voltages. I just found a photo of yet another transformer type used by APT sometime during production, so that makes 3 different types used that I'm now aware of.
The toroid I used is a Triad VPM36-690 medical grade - 18-0-18, 25VA with dual secondaries, tied to provide +/-/gnd.
The EI I used is a Hammond 167J36 - 18-0-18, 36VA center tapped. This was a better choice, because it has higher VA, and fits in the space better (not fitting a donut into a square box). It needed a tiny bit fitting, but was quite close. It has a thicker core and windings than the original.
Rectified voltages are obviously higher - around 24VDC, which provides good headroom for the Vregs.
The toroid I used is a Triad VPM36-690 medical grade - 18-0-18, 25VA with dual secondaries, tied to provide +/-/gnd.
The EI I used is a Hammond 167J36 - 18-0-18, 36VA center tapped. This was a better choice, because it has higher VA, and fits in the space better (not fitting a donut into a square box). It needed a tiny bit fitting, but was quite close. It has a thicker core and windings than the original.
Rectified voltages are obviously higher - around 24VDC, which provides good headroom for the Vregs.
This came up on my front page recommendations on aliexpress. Could they assist with our muting circuit, possibly bypassing it? I don't quite follow what they're saying, except that it is triggered on the falling edge of an input (approx 5v to 0v)
And the price, $0.57 TOTAL for 10 pieces, plus $2.18 shipping. insane
US $0.06 54%OFF | 10Pcs multi-function delay trigger timing chip module timer IC timing 2s - 1000h
And the price, $0.57 TOTAL for 10 pieces, plus $2.18 shipping. insane
US $0.06 54%OFF | 10Pcs multi-function delay trigger timing chip module timer IC timing 2s - 1000h
Thanks for asking! I'm still waiting on parts from Mouser that should arrive tomorrow. I'll post back with the outcome.
Sorry for not checking back in with the thread this week. I didn't have email notifications turned on for this thread, and I've been knee deep in two Sansui AU-717 overhaul projects. I'll definitely let people know what I end up doing. The plan for now is to revert to the old mute circuit. If that doesn't work then I'll put the old op amps back in (I used DIP sockets so that's simple) and see if that works out. At that point I'll make the call on a new transformer or leave it as is.
As (I think) I mentioned before, the 'upgraded' mute circuit had 24V connected directly to one end of the relay coil. The other end of the coil was connected to the new muting circuit on the floating PCB, which then continued to the mute switch and finally to 0V. The strange part was that this PCB also had its own jumpers connecting to +18V and +16V. These lower voltage rails were the ones I noticed sagging (down 200mV) when the switch assembly board was connected.
The interesting detail was that the +24V didn't sag when the switch assembly was connected. What happened was that the low side of the coil connected to the mute circuit wasn't able to as effectively pull down the voltage. So, as an example, when the mute circuit was working (switch assy. disconnected), the relay coil leads would measure +24V and +11V (relative to GND). When the mute switch wasn't working (switch assy. connected and +16/+18 rails sagging), the coil leads would measure +24V and +14V (again, relative to GND). So the voltage across the coil dropped even though it was fed +24V in both cases. So it seems like the sag on the +16/+18 rails kept the transistors in the mute circuit from opening fully.
The original mute circuit (if I'm reading it correctly) doesn't rely on the +16V or +18V rails, and just connects +24V to one end of the coil, and uses +24V to power the timer circuit. We'll see!
Not necessarily "better", but they'll work equally well. Almost any general purpose PNP/NPN would work. The delay circuit is not critical, and completely independent of the signal circuits, so no need for "audio grade" parts or tight tolerances or particular transistors.
4401/03 produce about 0.2-0.3 seconds longer delay, and about 1-2 mA more coil current.
Correct. The original circuit does not draw from the regulated supplies, so doesn't take away from the opamp power. The coil draws from the rectified but unregulated +24v. The delay circuit draws directly from the transformer secondary thru a single diode, for 1/2 wave rectification.
For grins, here's a sim of R90 at 4 values: 1k, 820, 560 and 390. Coil "hold" current increases as R90 value decreases.
When you get the "upgrade" coil board swapped out, please try to draw a schematic of how it's wired.
I gave it a shot this evening with a free online drawing tool. I'm sure I screwed something up, but maybe not...
The Apt Holman was one of the best measuring preamps of all time per the Audio Science Review review of one:
"Conclusions
The owner who sent me this has a knack for finding these past jewels with excellent engineering. Their performance post refurbishing is extreme good even by today's standard which sadly has sled backward. Hopefully no one sheds a tear next time I complain about an audio product producing SINAD of 70 dB and such which this level of performance was achieved years ago."
Also, the way the manual is written is like nothing else I've seen. It's pretty much an open book, describing the purpose and function of each part of the circuit. It's more like a white paper. The company only made 3 products as far as I can tell (Apt Holman, Apt Holman v2(?), Apt Amp 1), so I don't buy the claim (not yours!) that the buyers were fans who overlooked the shortcomings because of reputation. It had a reputation because was reviewed extremely positively as far as I can tell (and still is).
I really like the looks of it. It looks like science gear or avionics controls. No BS, no cursive writing, no silver bling, no self promotion ("direct hyper coupled micro drive blah blah". Just the name of the company, the model name, and the control labels.
It's the same reason I really like the look of NAD Monitor Series stuff from the 90s'.
Thanks for trying, but yeah, it doesn't quite "compute". Sim results show no delay, and coil current remains 0.
Have you removed the board already? If you can get a photo square on from the top, and another from the bottom, I can superimpose them to figure out the paths.
@seanc: here you go:
The transistor at the top is 2N3906. The one on the lower left side is 2N3904. The one on the bottom left is unmarked but the PCB says (2SC)1345.
Note that some of the jumpers to the main PCB are visible and are pointing toward the camera. Perspective makes these jumpers look like connections that don't actually exist.
I tried again with the drawing tool and saw mistakes in my earlier drawing. Also, I think when I refreshed the page at one point it reset some of the component values.
You're better off going off of the photos, but here's my second attempt. Looks way different.
The transistor at the top is 2N3906. The one on the lower left side is 2N3904. The one on the bottom left is unmarked but the PCB says (2SC)1345.
Note that some of the jumpers to the main PCB are visible and are pointing toward the camera. Perspective makes these jumpers look like connections that don't actually exist.
I tried again with the drawing tool and saw mistakes in my earlier drawing. Also, I think when I refreshed the page at one point it reset some of the component values.
You're better off going off of the photos, but here's my second attempt. Looks way different.