Adding current buffer to pre-amp
#16
Hi Mike, not sure what you mean here. It runs in class A up to 2W and above that either the top or the bottom transistors start to turn off. At powers less than 2W, the distortion components are < -80dB open loop, so will easily beat a "pure" class A.
#17
Brian,
What configuration do you call 'pure' class A - just curious as to what we are comparing with what ?
OK got your Diamond config as one of them (which I do like) but wondering what the alternative is that you have in mind..
Mike,
What is your 'pure' class A that you are referring too? Wondering if its the same as Brians..
James
What configuration do you call 'pure' class A - just curious as to what we are comparing with what ?
OK got your Diamond config as one of them (which I do like) but wondering what the alternative is that you have in mind..
Mike,
What is your 'pure' class A that you are referring too? Wondering if its the same as Brians..
James
#18
Hi James, I was assuming either an emitter follower or a common emitter with a constant current source, but you are correct, I didn't ask
Just to clarify, the distortion I quoted of < -80dB is without the opamp.
Brian
Just to clarify, the distortion I quoted of < -80dB is without the opamp.
Brian
#19
Brian,
I would make the same assumption but I thought we ought to check.
Nice figures on the Diamond beastie! I know its a horrible question but how does it sound? Are you pleased with it aurally - I guess you are with its engineering!
James
I would make the same assumption but I thought we ought to check.
Nice figures on the Diamond beastie! I know its a horrible question but how does it sound? Are you pleased with it aurally - I guess you are with its engineering!
James
#21
Hi James and Mike, not finished it yet. Built one board (I need four) and started the power supply. It's at this point in a project that lots of other things come in and take all of your free time
Anyway, hoping to do a bit tonight and get some volts on the board for some preliminary tests.
Found out that the resistor I wound the choke around is magnetic, so that will have to be changed too. I can't find non-magnetic resistors, so I'll find a former and use that.
Brian
Anyway, hoping to do a bit tonight and get some volts on the board for some preliminary tests.
Found out that the resistor I wound the choke around is magnetic, so that will have to be changed too. I can't find non-magnetic resistors, so I'll find a former and use that.
Brian
- Mike H
- Amstrad Tower of Power
- Posts: 20157
- Joined: Sat Oct 04, 2008 5:38 pm
- Location: The Fens
- Contact:
#22
Have to confess not heard of a diamond output stage until this topic so I looked it up and found this:
http://www.google.co.uk/patents/US20050264358
Spookily almost identical drawing:
http://patentimages.storage.googleapis. ... D00000.png
So looks like I'm torkin' bollox again
http://www.google.co.uk/patents/US20050264358
Spookily almost identical drawing:
http://patentimages.storage.googleapis. ... D00000.png
So looks like I'm torkin' bollox again
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
#23
I hadn't heard of it either Mike. Found this: http://www.audiodesignguide.com/my/Hot_Follower2.GIF and this: http://www.seekic.com/circuit_diagram/B ... rcuit.html
The second link shows a BUF634 buffer being used, but you'll see that the total schematic is what I'm building, but my buffer is discrete. The problem I have is that no buffers or power opamps seem to exist that are decent quality and about 2 watts output.
Brian
The second link shows a BUF634 buffer being used, but you'll see that the total schematic is what I'm building, but my buffer is discrete. The problem I have is that no buffers or power opamps seem to exist that are decent quality and about 2 watts output.
Brian
#24
Here's a full circuit so far and is the one I'm building.
If you are interested in LTSpice, look at the way variables are used to get 2^n samples for the FFT - gives much better resolution while allowing the frequency etc. to be easily changed. It's not my idea, I copied it
If you are interested in LTSpice, look at the way variables are used to get 2^n samples for the FFT - gives much better resolution while allowing the frequency etc. to be easily changed. It's not my idea, I copied it
- Attachments
-
- Diamond Output 7.asc
- (9.56 KiB) Downloaded 255 times
-
- zetex.zip
- (24.28 KiB) Downloaded 221 times
#26
You're welcome James.
Well, a bout of insomnia meant I could get the power supply started, finished, tested, mounted in a box with an on/off switch, the washing done and the apartment tidy - yes it was that bad. Still need to iron a shirt for tomorrow though...
Anyway, got the first buffer up and running and made a new invention. I'm going to call it the Colpitts Oscillator. Wanted to put 100 ohm in series with the bases of the first transistors, but the nearest I could find was 1k - the joys of moving house. It shut it up anyway.
Seems OK driving resistors, flat to about 10Mhz, rises slightly at about 12 MHz, then drops off nicely. Square waves are slightly ringy, probably related to the peak at 12MHz, but the input will be limited by a first order RC at 50kHz to 100kHz anyway, so I'm not going to worry about it. This all seems excessive, but it sits in the opamp feedback loop, so it's important that it doesn't introduce too much phase shift or the whole thing will oscillate.
Didn't realise how warm the output transistors would get, but I should have realised that class A without heatsinks would get a bit warm. They don't evaporate spit, so they have passed the traditional "lick test", but theoretically get to about 90 deg C, and touching them (ouch), I would say that is about right. The thermal stability is good with the input transistor sandwiched between two of the output transistors, the bias starts at 32mA per transistor (320mA total), and quickly settles to 27mA per transistor (270mA total). Balance between the output transistors is better than +/-10% so all in all, not a bad result.
Still no idea how it sounds though - need to build 3 more for that
Brian
Well, a bout of insomnia meant I could get the power supply started, finished, tested, mounted in a box with an on/off switch, the washing done and the apartment tidy - yes it was that bad. Still need to iron a shirt for tomorrow though...
Anyway, got the first buffer up and running and made a new invention. I'm going to call it the Colpitts Oscillator. Wanted to put 100 ohm in series with the bases of the first transistors, but the nearest I could find was 1k - the joys of moving house. It shut it up anyway.
Seems OK driving resistors, flat to about 10Mhz, rises slightly at about 12 MHz, then drops off nicely. Square waves are slightly ringy, probably related to the peak at 12MHz, but the input will be limited by a first order RC at 50kHz to 100kHz anyway, so I'm not going to worry about it. This all seems excessive, but it sits in the opamp feedback loop, so it's important that it doesn't introduce too much phase shift or the whole thing will oscillate.
Didn't realise how warm the output transistors would get, but I should have realised that class A without heatsinks would get a bit warm. They don't evaporate spit, so they have passed the traditional "lick test", but theoretically get to about 90 deg C, and touching them (ouch), I would say that is about right. The thermal stability is good with the input transistor sandwiched between two of the output transistors, the bias starts at 32mA per transistor (320mA total), and quickly settles to 27mA per transistor (270mA total). Balance between the output transistors is better than +/-10% so all in all, not a bad result.
Still no idea how it sounds though - need to build 3 more for that
Brian
#28
Hi Mike,
Mono would be ok, but still got to build another for that. I have a two way active crossover, and the buffer is going in the feedback loop of the last opamp in the crossover. I think it will be better to build four then I can completely dismantle my current system then go from there.
Mono would be ok, but still got to build another for that. I have a two way active crossover, and the buffer is going in the feedback loop of the last opamp in the crossover. I think it will be better to build four then I can completely dismantle my current system then go from there.
#29
Well, curiosity got the better of me. I dismantled a perfectly good stereo, and am now the proud owner of a mono - add that to your list DTB
I could go on about clean highs, earth shaking lows etc., but I think the middle is where it's at. It does sound like it goes on forever top and bottom, but the middle is sublime. It's a bit early yet, I need to make two more buffers to get back to stereo, and it's not a five minute job, but a couple of late nights should do it. The filtering all seems transparent (flat beyond 40kHz), and I haven't heard any pops and crackles yet, so another plus.
Brian
I could go on about clean highs, earth shaking lows etc., but I think the middle is where it's at. It does sound like it goes on forever top and bottom, but the middle is sublime. It's a bit early yet, I need to make two more buffers to get back to stereo, and it's not a five minute job, but a couple of late nights should do it. The filtering all seems transparent (flat beyond 40kHz), and I haven't heard any pops and crackles yet, so another plus.
Brian
- pre65
- Amstrad Tower of Power
- Posts: 21373
- Joined: Wed Aug 22, 2007 11:13 pm
- Location: North Essex/Suffolk border.
#30
Well done Brian.
I don't know enough to contribute anything positive but it makes good reading.
I don't know enough to contribute anything positive but it makes good reading.
The only thing necessary for the triumph of evil is for good men to do nothing.
Edmund Burke
G-Popz THE easy listening connoisseur. (Philip)
Edmund Burke
G-Popz THE easy listening connoisseur. (Philip)