Here is my build. Currently, I am using it as the main amp as it is able to drive my 8ohm FR speakers easily.
Whammy gets you head, pre, and power amp in a single package 😛
Whammy gets you head, pre, and power amp in a single package 😛
Hello, a bit at a time I managed to read through the whole thread, took some notes and have some questions. I’m a sophomore which means extra confusion, so bear with me:
I understand the three regulator options, I read in another DIYA post (Wayne’s prototype thread) to leave off the 1K R10/14 (and the cap across it) if using LED elevated rails or the 15V naked reg option, IOW use the 1K only for the 24V rail pre-amp option. However, I read the following from a Whammy post on AudioScienceReview:
"I think you actually want to keep R10 and R14, however, to bias the LED. I think the choice is to use the LEDs D5/D6 or R9/R13 in the ground leg but R10/R14 should always be there. Without them, it is just the regulators' trickle bias (quiescent) current through the LEDs, which strikes me as not terribly stable."
Makes sense to me as other regulator set ups I’ve seen had a resistor and cap btwn ground and out. So, for the naked or LED version keep or leave the 1K (and cap?) on the reg output?
-------
Speaking of regulators In some places, like around a 317, it is often better to use an old style general purpose e-cap with high-ish ESR to prevent oscillation, some designers expected the cap's ESR in their design, replacing old caps with snazzy super long life, ultra low ESR caps made for modern day switching power supplies can cause trouble in some analog circuits. If there any place in this circuit to avoid modern high spec caps and use a general purpose instead?
----
The Whammy originally spec'd a 22Vac/25VA tranny as Wayne wanted to allow use as a preamp and maybe use of high power op-amps plus a 22Vac was what he had on hand. There is repeated confusion on the thread whether a 15Vac or 18Vac is the best for cooler running as an HPA only. My math says 18Vac after rectification will force the regs to drop about 10V and 15Vac to drop 5V, 5V is more than enough for the regulator to function properly, when I want heat and wasted power I build tube amps 🙂
Is 18V recommended as the best alternative to 22Vac only because it gives a bit more flexibility with op-amps or does a 15Vac tranny not allow the LED elevation to work properly? Or, and what concerns me, is it because when driving high impedance phones the power supply won't have enough swing and will sag with a 15V tranny, and maybe not even with an 18V tranny? I note this comment from mod 6L6 on the prototype thread:
"Strictly speaking, the OPA2604 is necessary only if you use 24V rails. (Useful if you have very high impedance headphones (600ohm) and want to listen at deafening levels..."
I plan on using the Whammy as an HPA only and with Beyer 600 ohm phones mostly, occasionally at headbanger levels. If 15Vac gets me everything I want - LED quieting and uncompromised performance for 600R phones, I'll go with 15Vac. I'm ordering up the tranny, what's the final word for 600 ohm phones: 15Vac, 18Vac, or 22Vac together with a higher powered opamp?
-----------
Was 1uF input cap chosen as it's plenty fine and/or size constraints or because going bigger say, 4.7uf or 10uf as is commonly seen in this position causes subsonic instability? I ask as I have some good Wima 4.7uF caps that will drop right in.
Talks all.
I understand the three regulator options, I read in another DIYA post (Wayne’s prototype thread) to leave off the 1K R10/14 (and the cap across it) if using LED elevated rails or the 15V naked reg option, IOW use the 1K only for the 24V rail pre-amp option. However, I read the following from a Whammy post on AudioScienceReview:
"I think you actually want to keep R10 and R14, however, to bias the LED. I think the choice is to use the LEDs D5/D6 or R9/R13 in the ground leg but R10/R14 should always be there. Without them, it is just the regulators' trickle bias (quiescent) current through the LEDs, which strikes me as not terribly stable."
Makes sense to me as other regulator set ups I’ve seen had a resistor and cap btwn ground and out. So, for the naked or LED version keep or leave the 1K (and cap?) on the reg output?
-------
Speaking of regulators In some places, like around a 317, it is often better to use an old style general purpose e-cap with high-ish ESR to prevent oscillation, some designers expected the cap's ESR in their design, replacing old caps with snazzy super long life, ultra low ESR caps made for modern day switching power supplies can cause trouble in some analog circuits. If there any place in this circuit to avoid modern high spec caps and use a general purpose instead?
----
The Whammy originally spec'd a 22Vac/25VA tranny as Wayne wanted to allow use as a preamp and maybe use of high power op-amps plus a 22Vac was what he had on hand. There is repeated confusion on the thread whether a 15Vac or 18Vac is the best for cooler running as an HPA only. My math says 18Vac after rectification will force the regs to drop about 10V and 15Vac to drop 5V, 5V is more than enough for the regulator to function properly, when I want heat and wasted power I build tube amps 🙂
Is 18V recommended as the best alternative to 22Vac only because it gives a bit more flexibility with op-amps or does a 15Vac tranny not allow the LED elevation to work properly? Or, and what concerns me, is it because when driving high impedance phones the power supply won't have enough swing and will sag with a 15V tranny, and maybe not even with an 18V tranny? I note this comment from mod 6L6 on the prototype thread:
"Strictly speaking, the OPA2604 is necessary only if you use 24V rails. (Useful if you have very high impedance headphones (600ohm) and want to listen at deafening levels..."
I plan on using the Whammy as an HPA only and with Beyer 600 ohm phones mostly, occasionally at headbanger levels. If 15Vac gets me everything I want - LED quieting and uncompromised performance for 600R phones, I'll go with 15Vac. I'm ordering up the tranny, what's the final word for 600 ohm phones: 15Vac, 18Vac, or 22Vac together with a higher powered opamp?
-----------
Was 1uF input cap chosen as it's plenty fine and/or size constraints or because going bigger say, 4.7uf or 10uf as is commonly seen in this position causes subsonic instability? I ask as I have some good Wima 4.7uF caps that will drop right in.
Talks all.
Last edited:
I would say that capacitors directly close to the regulator output should not have too lowESR.
I think that doing that further away; like for example right next to the output stage mosfet (or definitely on OpAmp rails) won't bring any risk of regulator instability given the additional serial resistance due to the PCB.
That being said; nothing is perfect, and you run the risk of creating a resonator with all those parasitic R/L/C.
-> Better check with a scope when experimenting 😉
I don't have experience on this, so I just followed the guide and didn't use them.
I wonder if they'd bring much on the table. When your output is too high, you also bias regulator's GND up, which in turn increases the output... Does not sound like a very stable/accurate mechanism to me. I have beefed up the LED capacitors; as I think it cannot hurt (moar is bettah)
Naked reg ? -> 15Vac
LED/Resistor red? -> 18Vac (15Vac might be enough but it does not leave much room on power hungry headphones)
I personally used 22Vac and use LED reg config; but this is too much dissipation on regulators for my taste; I'd better have used 18Vac (not available at the time).
It depends on which beyer you have, but 15Vac+Naked reg should already be plenty.
OpAmp power/current capability does not really matter as they are only lightly loaded by the gates of the output transistors.
You may look at:
https://diyaudioheaven.wordpress.com/tutorials/power-impedance-etc/
's HOW much power you actually need for your headphone(s)?
And see whether you'd be limited in voltage ; or current (power)
0 risk in using higher values in my view.
C26/C27 anyway already creates a high pass.
Personally, I got rid of those input caps because all my sources are DC-free.
No. The output voltage from the opamp can swing close to the rails but the FET output stage has significant limitations due to the gate-source voltage of the FET's
So a -/+15 volt supply will see the final amp output deliver around lets say -/+8 volts to the load. The opamp output will be closer to -/+ 12 volts or more under those conditions.
So a -/+15 volt supply will see the final amp output deliver around lets say -/+8 volts to the load. The opamp output will be closer to -/+ 12 volts or more under those conditions.
I'm using 18VAC xformers with 15V Sparkos regs. So, I'm guessing the max voltage swing my amp can output is about +/-8V?
Kinda wish I would've gone with higher rails now, just for that extra bit of headroom. 😛
Do you listen at 110db? Or do you have incredibly inefficient headphones? Swinging 16v across any normal headphone will produce SPLs into the range of instant hearing damage.
Please. No, music sounds lifeless at any level below 120dB. That's why I now use the SissySIT's speaker taps directly for casual listening. Attenuation is for Sissies.
😜
Nah, I don't need higher rails. I think the pot position at noon is the max I could listen at with my HD650's or HD800S. It's just the power that comes along with knowing "it's there if you want it."
Also, audiofool myths that plague my mind with "higher rails are better" and "more headroom equals better transients." Even if I know they're misconstrued by audiophiles.
I'd have to check again, but I think they're around 13.5-14V under load. So, if I'm getting 10V out of this amp, that's pretty solid lol.
I'm not going to retrofit my Whammy. It works and sounds great. I was just thinking that if I were to build it over again, I'd probably try playing with 18V regs -- just for fun >: )
Excellent build technique Mr Scranton on post 4639 and the simple shaft connection for the pot too - this should serve as an example of "smart/better engineering"
Your load is 600 ohms. With +15 / -15V regs (and even with that output stage biased at around 60mA), I wouldn't expect that much of a voltage drop...??
You're right. I misremembered. I just checked and it idles at around +/-14.5VDC at the opamp pins.
I have just completed my Whammy. So impressed with the great design and rapid delivery from diyaudio store. It is grounded properly, my volume pot is grounded to chassis. I have checked all resistors for correct value and correct position. I have the LED voltage reference and a 4850DD op amp. My power supply voltages are both in 16v range. The left channel DC offset (measured with no input or output connected is nearly double the right channel. (12 mv vs 5.8 mv). It is louder in the left channel and after leaving it on for about30 minutes the output devices for the left channel are much warmer than the right. It then distorts badly. I have also tried a LM833N which shows even bigger difference between the channels. I have rechecked all solder joints. Where should I start troubleshooting next?
Attachments
Can you take closer pictures of the regulators from the top and bottom of the board? Have you double-checked the resistor values? You did a nice job of leaving the resistor codes on top, but I cannot read all of the values with the photo you attached.
Chris, thank you for the prompt reply. I reflowed all the solder on the bottom of the PCB and now the difference is much smaller (7mv vs 5 mv). I have pu the PCB back in the case and will remove it and take the photos over the weekend. Cheers, Pieter
1. Check all resistor values.
2. Solder Opamp directly into the board. Do not use dip adapter.
3. Check the Alps RK27. It might have channel imbalance.