Hello guys, i`m trying to put together a headphone amp by using Tangentsoft`s approach ( http://tangentsoft.net/audio/cmoy-tutorial/misc/cmoy-tangent-sch.pdf ) but i fail to understand how to wire the virtual ground. I just can`t figure out where to connect the wires. I made a Fritzing design of where i am now. If anyone could tell me what to do with the wires hanging around in the picture( https://dl.dropboxusercontent.com/u/45794359/amp_bb.jpg ) i would really appreciate it. If someone finds any other mistakes please feel free to point them out.
I am also using a virtual ground, but on a much larger scale in an effort to 'rehab' a dynaco ST-120. (I also did one using 'updatemy dynaco' circuits, but for this one, I want to use chipamp.com's LM3886 boards, thus the need for bipolar ps.) So, I created a simple regulated PS using a virtual ground created off the diode bridge with two big resistors followed by two big capacitors followed by a simple regulator using IR240/9240. (I am traveling and do not have the ability to scan and post a schmatic, sorry.)
Anyway, my question is: how to manage the grounds? The ST-120 does not have a third wire power cord, but I'm not sure that makes a difference and I may add one. The chipamp board connects Input gnd, Output gnd, and Power gnd (+ and -) to the same fairly large ground plane. My plan would be to run twisted pairs for power +/gnd, powr -/gnd, input +/gnd, and output +/gnd, all isolated from chassis. The question then becomes WHAT/WHERE do I connect to chassis? My thought would be to use a heavy wire from the virtual ground on the PS board to chassis with no 'loop breaker'. Since signal and power gnds are connected on the chipamp board, I cannot 'lift' the signal gnd from the chassis unless I 'lift' the PS as well which strikes me as a bad idea. Of course in its present form, the chassis is not, itself "grounded" since it only has a two wire power cord.
Coments/advice welcomed!!
Anyway, my question is: how to manage the grounds? The ST-120 does not have a third wire power cord, but I'm not sure that makes a difference and I may add one. The chipamp board connects Input gnd, Output gnd, and Power gnd (+ and -) to the same fairly large ground plane. My plan would be to run twisted pairs for power +/gnd, powr -/gnd, input +/gnd, and output +/gnd, all isolated from chassis. The question then becomes WHAT/WHERE do I connect to chassis? My thought would be to use a heavy wire from the virtual ground on the PS board to chassis with no 'loop breaker'. Since signal and power gnds are connected on the chipamp board, I cannot 'lift' the signal gnd from the chassis unless I 'lift' the PS as well which strikes me as a bad idea. Of course in its present form, the chassis is not, itself "grounded" since it only has a two wire power cord.
Coments/advice welcomed!!
I'm afraid you didn't read my post correctly. I am not using batteries. I am using the transformer from a dynaco ST-120 which does not have a CT, thus the need for a virtual ground.
I'm not convinced that it's a good idea to run a chip amp with a "virtual" ground. I'd be willing to bet my lunch money that it would introduce stability and/or distortion problems when pressed into such an application.
If you must use the Dynaco transformer, the easier solution (which would probably give the best results) would be to modify the circuit for single sided supply. You still have the big output caps from the Dynaco, right?
Maybe a better solution would be to find a center tapped transformer from an old amplifier or receiver. I have one right now out of an old Pioneer receiver that I swear must have just been replaced before the unit was left at the electronics recycling dumpster. The outputs were brand new too, but fried. Anyway, some dumpster diving, garage sales, thrift stores, etc might turn up a better transformer.
There's people that collect those old Dynacos. I know a guy that would have a fit if he knew what you were doing.
If you must use the Dynaco transformer, the easier solution (which would probably give the best results) would be to modify the circuit for single sided supply. You still have the big output caps from the Dynaco, right?
Maybe a better solution would be to find a center tapped transformer from an old amplifier or receiver. I have one right now out of an old Pioneer receiver that I swear must have just been replaced before the unit was left at the electronics recycling dumpster. The outputs were brand new too, but fried. Anyway, some dumpster diving, garage sales, thrift stores, etc might turn up a better transformer.
There's people that collect those old Dynacos. I know a guy that would have a fit if he knew what you were doing.
You can use the center point of the batteries as ground, Nothing wrong with that at all and should be how it is done.
The only exception is that if the batteries don't discharge equally than one may have a lower voltage then the other one and this can cause asymmetrical distortions and/or clipping of the waveform.
Using a the resistors assures that the ground would be exactly half of the supply voltage even as the voltage gets lowered when the batteries are draining.
This method is only good if you are only using a few stages of opamps, If you are using many more opamp's then it is advisable to use a Virtual Ground buffer (for a single ended supply) to assure a low impedance ground system for all of the common opamp stages.
This has been covered in great detail in many many and recent threads.
FWIW
jer
P.S. The above is for general use in opamp circuit's and not meant use for high current situations such as in a power amplifier.
In cases such as that a proper Bipolar Power Supply should be used.
The only exception is that if the batteries don't discharge equally than one may have a lower voltage then the other one and this can cause asymmetrical distortions and/or clipping of the waveform.
Using a the resistors assures that the ground would be exactly half of the supply voltage even as the voltage gets lowered when the batteries are draining.
This method is only good if you are only using a few stages of opamps, If you are using many more opamp's then it is advisable to use a Virtual Ground buffer (for a single ended supply) to assure a low impedance ground system for all of the common opamp stages.
This has been covered in great detail in many many and recent threads.
FWIW
jer
P.S. The above is for general use in opamp circuit's and not meant use for high current situations such as in a power amplifier.
In cases such as that a proper Bipolar Power Supply should be used.
Last edited:
I would like to see the schematic of this virtual ground circuit.
For this application, it would have to be large and fast. I picture a fast, precision voltage divider driving a high impedance input servo circuit. The buffer portion would have to be super fast and able to deliver large current, or else distortion would be introduced.
This seems impractical to me. I'd rather source another transformer, whether that means scrounging second hand stores and garage sales or dumpster diving, or else just ordering s new transformer.
I still think that a single ended circuit might be a cheap and practical solution. Do you still have the big chassis mounted caps? With them, the original transformer, and a handful of extra parts, you're good to go.
The datasheet should have an example of a single sided circuit configuration.
For this application, it would have to be large and fast. I picture a fast, precision voltage divider driving a high impedance input servo circuit. The buffer portion would have to be super fast and able to deliver large current, or else distortion would be introduced.
This seems impractical to me. I'd rather source another transformer, whether that means scrounging second hand stores and garage sales or dumpster diving, or else just ordering s new transformer.
I still think that a single ended circuit might be a cheap and practical solution. Do you still have the big chassis mounted caps? With them, the original transformer, and a handful of extra parts, you're good to go.
The datasheet should have an example of a single sided circuit configuration.
In case you were asking a real--not rhetorical--question, nothing's wrong with it, it is ground.
To the electricians who wired your house, "ground" is literally that, it's the planet earth. There's a ten-foot rod driven into the earth outside and that's ground.
But in electronics "ground" is arbitrary. It's a point you define as zero volts, and then all voltage measurements are relative to that point.
.
Last edited:
Here are just a few examples,
http://www.diyaudio.com/forums/anal...-noob-preamp-single-supply-3.html#post3853760
http://www.diyaudio.com/forums/soli...le-supply-preamp-amp-circuit.html#post3283860
and even more here,
https://www.google.com/search?q=Vir...yQSi4IHoBw&ved=0CCsQrQIoBDAA&biw=1152&bih=744
jer
Neither of those schemes looks very promising. You can toss any schemes for preamps or headphone amps out right away. No way Jose.
Did you read the thread in the second link? It doesn't look like it worked very well under a dynamic load.
Did you read the links? You are not the first one to think of this scheme. It is a good idea but impractical for power amplifiers in practice. The complexity of a buffer that would actually work in this application would far exceed any 3886 circuit. You would in essence be building a super high speed, super high precision, high current DC buffer.
Or you could configure the 3886 circuit for single ended supply, and employ the old big boy output capacitors. I guarantee you it will work way better and sound way better than any virtual ground scheme you can come up with. And as a bonus you won't need speaker protection. If I remember correctly, the Dynaco amplifiers came with an inductor wound around the output capacitors. Add a resistor and you have a Thiele network.
Last edited:
I won't claim to have read EVERY article on virtual grounds, but I have certainly read a lot. I have come up with a virtual ground that I am comfortable with (again, sorry, but I do not currently have acces to a scanner. I will post schematic when I can.) Anyway, the question I asked was NOT how to design a virtual ground or which design to use, but HOW/WHERE to connect it both to the chassis and to the specific chipamp.com board. On the chipamp board, there are attachment points for FOUR grounds (V+/G, V-/G, Sig In/G, Sig Out/G) all of which are joined in a large ground plane. My plan is to use ALL FOUR , using twisted pairs. Now at this point all the "grounds" are accounted for, but NOTHING is connected to the chassis or EARTH. Where/how do I do that?
Here ya go. Note that the correct term for a connection made to a building's grounding system is "bond," so that's the term I use here.
All bonds ultimately connect to the building's grounding rod driven into the earth that I mentioned before. Quite distinct from the "ground" used in electronic circuits.
In the illustration, the chassis bond point (one and one only) is shown as being very near the audio input, and connected to the input shield. This is correct.
.
Attachments
Last edited:
Your dangerously misinformed statement could get somebody killed. Since you obviously know nothing about how buildings are wired, what possible reason can you have for saying such a thing?!
.
Last edited:
Does point 'A' also connect to the chassis? If yes, doesn't that create a ground loop? If no, what does the ground symbol refer to?
The green triangle symbol is a mechanical and permanent connection to Chassis.
The triangle symbol next to point A is a mechanical connection to Chassis.
The first is to comply with:
the Protective Earth must be permanently connected to Chassis.
The second is to comply with:
All exposed conductive parts must be connected to the (protected) Chassis.
Note that this diagram makes the same mistake as described in post36
http://www.diyaudio.com/forums/chip...-grounding-methods-designs-2.html#post4242368
The screen if also acting as Signal Return MUST stay close coupled to the Signal Flow over the whole route of the two wire connection.
If a screened twisted pair was used as the signal input then the two twisted wires carry the signal into and out of the amplifier. The screen is then correctly tied to Chassis at the entry aperture.
Bent's diag omits the necessary Close Coupled Return wire.
Last edited:
Ohh, I am sorry My bad, I thought this thread was about the little Cmoy amplifier.
This subject should have been a new thread.
I personally would never consider a virtual ground system for a power amplifier.
In such a case were I had no choice of a center tap for a proper Bipolar supply i would choose a BTL configuration for the amplifier.
Or a Voltage doubler can be used with some very large capacitors to form a Bipolar supply's I do this with opamp's all of the time.
I just never wasted my time doing it with a large power amplifier and always did it the proper way.
Cheers!!
jer
Last edited:
Looking back at the original Dynaco wiring, I see that the shield of the RCA plug was physically connected to the chassis, but it also had a twisted pair going to the amp pc board. This would satisfy both requirements, yes? In addition, there was a star ground where the power supply ground AND the signal ground was connected. This is the wiring I was thinking of emulating, but it seems to me that it creates a ground loop because the signal ground is connected to the chassis at TWO points (input socket and star). I guess if it was good enough for Hafler (and it obviously worked without problems), it should be good enough for me. Just trying to understand the potential problems before I make decisions that will be hard to undo.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.