Need help with PS for CMoy amp.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello,
I am a beginner in circuit design and I need help in power supply design for this amp. Since my house is covered with snow, I live in Northern VA, I would like to re-build my CMoy amp with slightly different input cap and carbon resistors. For an input cap I want to experiment with polycarbonate, polypropylene, and old paper in oil ones ( all 1 uF). Since it is not portable design I would like to build two power supplies: one battery with TLE chip and one regular with the use of transformer which I already have and I hope good for the purpose. The amp uses OPA134PA chips. I will use it with three Sennheiser phones: HD 455 (52 Ohm), HD 480II (I am not sure impedance, but it should be at least 32 ohms), and HD202 (32 ohms).
The transformer, QLP0540S, is from old Technics cassette deck RS-271US. I do not know what type of the trafo it is, but it looks almost like minature version of cut-core one. The trafo is design for multiple input voltages: 117, 220, 100, 125, and 240. In current setup it is set to 117 V. Output is (in order of marked outputs from the left); 24V 9V 0 5V 9V 24V. I still see wires coming from the output connectors as they were used in the deck: 0-5V, 9V-9V (this is probably for motor?), and 24V-0-24V.
What I would like to do is to try to connect them 9V-0-9V. I hope I can do that, what would give me +/- 12.69 V after rectification.
So, the first question would be, can I use that trafo for the purpose? I do not think the OPA134 needs a lot of current.
I want to build unregulated PS, the next question is what capacitance should I use?. All power supply examples fo chip amps, which I found on Headwize website, are regulated.
Can someone assist me, please?
 

Attachments

  • Trafo1.jpg
    Trafo1.jpg
    55.7 KB · Views: 248
  • Trafo2.jpg
    Trafo2.jpg
    71.1 KB · Views: 245
  • Trafo3.jpg
    Trafo3.jpg
    52.5 KB · Views: 253
Just to say more what I want to do is: except experimentation with input cap, I want to use carbon composition resistors. Also bypass supply pins with combination of 100uF/ 0.1uF ceramic capacitors. Also to add 0.01uF plastic cap between supply pins under the chip. So, as you see it is a sort of experiment. I like the sound of the orginal CMoy amp with gain of 11 from Headwize. This idesign is to slightly ....change it character.
 
So, the first question would be, can I use that trafo for the purpose?
The 9-0-9 V were probably there for the op amps in the tape deck, so use that trafo.

I want to build unregulated PS, the next question is what capacitance should I use?. All power supply examples fo chip amps, which I found on Headwize website, are regulated.
The reason is that you need big amounts of capacitance to suppress the 100 or 120 Hz hum you get after rectification. Regulated power supplies need much smaller capacitors, which often turns out to be less expensive and smaller than an unregulated power supply for the power levels of op amps. And they give better results. While the op amp itself usually has pretty good PSRR, you cannot always avoid ground loops. And a headphone will make them heard. Regulated power supplies will suppress that hum more efficiently than unregulated ones.

I would estimate that you need at least 6800 µF to avoid hum issues on an unregulated supply with 2 x 32 Ohm load. For a regulated power supply it would probably be sufficient to use 150 µF plus the small capacitors that are specified in the corresponding datasheet.

On the other hand it depends on the listening level. And none of your Sennheisers is extremely revealing, so you may get away with much less capacitance without noticing much of a difference.

If you want to stick with the unregulated PSU, get a bunch of 1000 µF capacitors with the right voltage rating, start with one per rail and add more until the hum is gone and you like the sound. Or until you find that you cannot kill the hum and have to go regulated. You can improve the performance by adding resistors in between the capacitors. The bigger the resistance, the better the filter, but the lower the rail voltage. Trial-&-Error is the way to go to find the best compromise for yourself.

You may find that the CMoy reaches its current limit easily with a 32 Ohm load. Oh, and the HD480 has 70 Ohms.
 
Hello Pacificblue,
Thanks for the response. I do have a lot of capacitors and I wanted to make it as simple as possible. I do not like the idea of a lot of experimentation to get rid of hum.
In this case I can use 24V connection and regulate down to 12 volts (I am not sure I can use 0 output with 9V one to get symetrical voltage. I will need to test that trafo. For sure I have 24-0-24.)). I am not sure I have negative voltage regulators though.
oh well I will check.
 
In this case I can use 24V connection and regulate down to 12 volts
The resulting heat dissipation may force you to heatsink the regulators. 24 V AC will give you ~33 V DC without even taking regulation and supply tolerance into account.
If you connect the transformer's 220 V or 240 V primary to mains, you should get something around 12 V AC from the 24 V output. That should be okay for a 12 V regulator.

I am not sure I can use 0 output with 9V one to get symetrical voltage. I will need to test that trafo.
You wrote, you have 24-9-0-5-9-24. If the transformer is not broken, you can use 9-0-9 to create a split power supply with 8 or 9 V regulators. You can also use a single supply and create a virtual ground at half the rail voltage. There are examples of CMoys that run from a single 9 V battery.
 
Administrator
Joined 2007
Paid Member
If you are on 120 ??? volts or so AC and the tranny has a 240v winding you could use that to reduce the outputs to around half their value. So the 24 0 24 would become 12 0 12.

Even though it's a low powered amp you need to pay regard to wiring and grounding. If you have "hum" which is a pure low frequency tone then it's more than likely stray pickup. If there is "harshness" to the buzz (harmonics) then it's a sign that the wiring is incorrect and that 100/120 hz (twice mains freq) is contaminating the grounds.
 
Hello,
Thank you all for the help. I did find voltage regulators !!! Those are L7812CV and L7912CV. I do have smaller heatsinks from older computers and monitors.
But, I will try to do what you suggest to use 240 Volt tap. The only question is about voltage drop at regulators. I am not sure what LXXXX series of regulators require at their input to maintain output voltage. Rectified 12 volt output will give 16.92 volts. I guess this is enough, but that 12 volts is being generated by part of transformer not design for 117V. How about supply of needed current, is it OK? I am asking because I am doing that for the first time and I even did not know you can connect transformer that way (I understand we can connect transformer design for higher voltage to lower one). From my point of view it is fine. PS enclosure would be smaller without heatsinks. But, if PS might have some problems I would rather use heatsinks.

And, as we already discussing it what is the best way to connect outboard PS to the amp housed in its own enclosure? What I have noticed the recommendation is to connect power ground to power supply box only. Do not connect amp enclosure to power ground, only connect signal grounds together and connect all to 0 wire which brings power from PS. Is that correct?
Thanks again !!!!
 
Administrator
Joined 2007
Paid Member
You are going to have to try it and see is the simple answer. The theory is sound, it's whether it behaves quite the same in practice... the regulation of the tranny for example is unknown... but at these currents it's a non issue, as an opamp uses milliamps only.
Input to the regulators should be fine at 16 vdc. What really matter is that the ripple component on the unregulated supply doesn't go below the drop out voltage for the reg but at milliamp currents again it's a non issue.
 
Administrator
Joined 2007
Paid Member
External PSU is fine if that's what you want.
If I were making this I would use a small transformer for what they cost, something like,

BLOCK|AVB3.2/2/12|TRANSFORMER, 3.2VA, 2 X 12V | CPC

and build it all as one.
Get the grounding correct and there should be zero hum from the PSU electrically. That means taking the ground connection (star) from a small spur from between the two main smoothing caps. That's then the only connection you use to the amp etc.
As to induced hum from the magnetic field of the tranny... at these signal levels it shouldn't be a problem. A bit of distance tranny to circuit works wonders and rotating the tranny to cancel the field helps to in some cases. At the impedances of the CMOY and the levels involved I would try all in one box tbh... and make it work :)

If you have "hum/buzz" and it's "harsh" sounding because of harmonics, then that points to a wiring problem. Pure low frequency hum is induced or caused by stray pickup.
 
Hi,
I have one more question about regulated PSU. I still would like to build external one and I am not sure where to install regulators. In majority designs it is recommended to install regulators as close to chip as possible. But, in my case would be easier to do it inside of the external PSU case. Does it matter for this particular project? What might be important info, I do not want to use long umbilical cord between PSU/Amp. For this purpose I want to use shielded computer power cord.
But, in the case of use of regulators in the amp case as a better solution, should I put any capacitors after bridge rectifiers (inside of PSU box)?
 
Administrator
Joined 2007
Paid Member
For your application it won't matter. Just remember to decouple the IC's (OpAmps) at their supply pins directly.

The supply from a regulator is "at it's best" directly on the regulator... once you feed that voltage down a wire or a length of PCB print that adds a small resistance, which means that load fluctuations "alter" that super stable voltage at the "load end" of the wire. In practice it's worse than that as the wire etc add inductance too so that the AC impedance of the supply is higher at the load end as well. So decouple well at the amp.

The good news is that for a CMOY it won't make one jot of difference :)
 
Mooly,
Thank you very much for the help !!!!!
I will do what you have suggested, regs close to chips. I wanted to bypass power pins at the chip anyway. In that case, should I put any smoothing caps in the PSU after rectifiers or it is not necessary?
 
Administrator
Joined 2007
Paid Member
You definitely use caps in the PSU lol.
The caps at the regs need not be large, they are more for stability. The problem with putting big caps "in and around" an amplifier (again yours is so low current so it's not a problem) is that any ripple, any disturbance due to the load drawing current, then that current is introduced via large caps into the signal ground causing an increase in distortion etc

So smoothing caps in the PSU, probably 1000uf at most for this low current application, then smallish caps at the regs, say 47uf, and something like a 10uf and a 0.1uf in parallel directly across the supply pins to the OpAmp.
 
Cmoy and Zobel

It took me two years to get back to construction of this amp. I do not want to pollute the forum with another thread about Cmoy amp, but after search I did not find an answer to the Zobel question. Maybe because this amp is usually made portable and additional parts are not desired. Anyway, after lurking through RJM site I have found his version of Cmoy. It has Zobel at the output with note it is recommended. I have not seen here anyone implementing Zobel in Cmoy. So, would using it improve the sound or improve stability with lower impedance headphones? Do you recommend it?
Thanks

This is URL to RJM site: http://phonoclone.com/diy-hclo.html
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.