Re: Re: Could a tube buffer be useful for a Non-Inverting Gainclone?
Joe, as an alternative to a buffer feeding a ICG, "kuei yang wang" (aka:Thorsten?) suggested a "Grounded Cathode" gainstage feeding the LM3875 configured as a zero gain IGC, in effect letting the influence of the tube stage set the overall "sonic flavor" of the circuit. This is similar to the path my own efforts have been following. I'd be interested in your comments, given your involvement with Allen W and Vacuum State Electronics.
Thanks in advance!
-Tom
Joe Rasmussen said:
Joe, as an alternative to a buffer feeding a ICG, "kuei yang wang" (aka:Thorsten?) suggested a "Grounded Cathode" gainstage feeding the LM3875 configured as a zero gain IGC, in effect letting the influence of the tube stage set the overall "sonic flavor" of the circuit. This is similar to the path my own efforts have been following. I'd be interested in your comments, given your involvement with Allen W and Vacuum State Electronics.
Thanks in advance!
-Tom
fdegrove said:Hi,
Any reason why you can't use the one for the IGC?
Or the one Moamps designed with the CCS in the tail, for instance?
Cheers,
Ok, earlier I implied my lack of sophistication. Now I'm forced to demonstrate it conclusively. Using Joe's buffer on a NICG can't be as simple as flipping the connections to the chip can it? Seems like one more reversal is needed to get feedback in phase?
Sheldon
fdegrove said:Hi,
I'm by no means an expert on gainclones but a buffer is a non-inverting circuit.
Basically you could put one in front of any amp if you'd wanted to.
Cheers,
I understand that in the general sense. Specifically, how would Joe's schematic change to connect to this? Or, what would the schematic be for combining Joe's basic buffer with this amp?:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=28743
More generally, what would the schematic (signal input to signal output) be for Joe's buffer, including the pot for use as a standalone buffer?
Sheldon
Sheldon said:
Basically it boils down to this: What changes are needed to make it a V-NIGC ?
A few changes in values are required but I have tried to keep them as similar as possible to VIGC. Keep in mind that this hasn't been built yet, but I don't foresee any real problems. But it probably makes sense for someone with previous experience, who has already done a VIGC using earlier values, make convertion and then report back.
An externally hosted image should be here but it was not working when we last tested it.
Again the usual recommendation to tweak the 1n3* value.
I will post this also on the main thread that started all this.
Joe R.
Re: Re: Re: Could a tube buffer be useful for a Non-Inverting Gainclone?
A few thing to say about it:
1) Thorsten's approach is completely different. Not only does the tube's added gain add 'flavour' but also the interface with loudspeaker load simulate higher output impedance of low or zero feedback tubes.
2) Complexity: I make no apology for the simplicity of the VIGC as I envisioned it. This was all about making it easy for almost anyone to try dabble with DIY tubes.
3) High Tension: The VIGC has SS levels of voltages, Thorsten's aka KYW has HT. That is a disincentive for those reluctant to tackle tubes.
4) The VIGC is a cut-down version of the JLTi. In that design I did intentionally not want to 'warm up' the sound. Indeed the end result shares some of the characteristics of my ILA (exotic tubes, neither SE nor really PP) in terms of soundstage and transparency and high (but not etched) resolution. That is the kind of sound that Allen and I have strived for, low colouration tube sound, with great refinement and truthfulness. Not all may want that but we have managed to get a lot of folks to come with us for the ride. Even the VIGC does a pretty good job in that respect, so it seems a lot do agree with us.
Joe R.
tecnofossil said:
Joe, as an alternative to a buffer feeding a ICG, "kuei yang wang" (aka:Thorsten?) suggested a "Grounded Cathode" gainstage feeding the LM3875 configured as a zero gain IGC, in effect letting the influence of the tube stage set the overall "sonic flavor" of the circuit. This is similar to the path my own efforts have been following. I'd be interested in your comments, given your involvement with Allen W and Vacuum State Electronics.
Thanks in advance!
-Tom
A few thing to say about it:
1) Thorsten's approach is completely different. Not only does the tube's added gain add 'flavour' but also the interface with loudspeaker load simulate higher output impedance of low or zero feedback tubes.
2) Complexity: I make no apology for the simplicity of the VIGC as I envisioned it. This was all about making it easy for almost anyone to try dabble with DIY tubes.
3) High Tension: The VIGC has SS levels of voltages, Thorsten's aka KYW has HT. That is a disincentive for those reluctant to tackle tubes.
4) The VIGC is a cut-down version of the JLTi. In that design I did intentionally not want to 'warm up' the sound. Indeed the end result shares some of the characteristics of my ILA (exotic tubes, neither SE nor really PP) in terms of soundstage and transparency and high (but not etched) resolution. That is the kind of sound that Allen and I have strived for, low colouration tube sound, with great refinement and truthfulness. Not all may want that but we have managed to get a lot of folks to come with us for the ride. Even the VIGC does a pretty good job in that respect, so it seems a lot do agree with us.
Joe R.
Sheldon said:
Just some additional construction thoughts:
1) This circuit is not strictly U-G (Unity Gain) @ DC, hence any DC imbalance at the input is likely to be amplified at the output as DC Offset.
2) To make it U-G, use a largish cap, but not electrolytic cap, should be inserted between 18K connected to the (-) input and ground. I would be inclined to use 10uF. But also inclined to avoid using it at all.
3) Thorsten's most recent IGC that I have seen does not look strictly not U-G either (the earlier one is as well as the VIGC). Low DC values and well balanced inputs may avoid using U-G cap. It is worth a try. In Thorsten's his gain is a lot lower and DC is potentially amplified less. The 820K/18K feedback ratio gives a gain of 46 (33dB) and this could be a problem without the additional U-G cap.
4) It might be a good idea, if no U-G cap is to be used, to put another 820K resistor in parallel with the input 18K, since in DC terms the 820K/18K feedback resistors are also in parallel (so both inputs sees 17K613 @ DC). The difference between 18K and 17K613 x 46 could be significant?
So, as said, this is yet to be constructed and sussed out. Best to be tried by a current experienced VIGC constructor?
Joe R.
An externally hosted image should be here but it was not working when we last tested it.
Single power supply
I would like to give this a try as well. I have now built several NI Gainclone "channels" and have been looking at Joe's tube buffer.
My question is- since my PS puts out about 31VDC into a load is there a way I can just split it and use it to power the valve as well? I realize the valve needs low current and so the high current going to the chips isn't suitable but is there a way to restrict the current without resorting to a separate transformer and bridge for the tube?
I would like to give this a try as well. I have now built several NI Gainclone "channels" and have been looking at Joe's tube buffer.
My question is- since my PS puts out about 31VDC into a load is there a way I can just split it and use it to power the valve as well? I realize the valve needs low current and so the high current going to the chips isn't suitable but is there a way to restrict the current without resorting to a separate transformer and bridge for the tube?
Re: Single power supply
I'm not an electronic expert, but I can give a very basic answer. Too large a current flow is not the issue, as the current is determined by the resistance of the tube circuit at a given voltage. The analogy for voltage is pressure in a water pipe. Current is then flow volume, which is a function of pressure and the size of the pipe. The chip is a pretty big pipe at high outputs, the tube buffer is a much smaller pipe. But they can both take the same pressure (voltage) and either could be connected to a power supply much larger than required, as long as the correct voltage is applied.
The reason for separate power supplies is to isolate the tube buffer from the voltage variation that will be caused by chip's demands on the power supply. The experts here can tell you all the bad stuff that can happen with the buffer and amplifier talking to each other via the power supply. The buffer is giving directions and doesn't want any argument or backtalk. As a neophyte, I'm guessing that there are ways to use the same supply transformer, but I bet the circuits necessary to give the buffer clean power from the amplifier chip supply would be more difficult to execute than just using a separate supply.
Sheldon
Sherman said:
I'm not an electronic expert, but I can give a very basic answer. Too large a current flow is not the issue, as the current is determined by the resistance of the tube circuit at a given voltage. The analogy for voltage is pressure in a water pipe. Current is then flow volume, which is a function of pressure and the size of the pipe. The chip is a pretty big pipe at high outputs, the tube buffer is a much smaller pipe. But they can both take the same pressure (voltage) and either could be connected to a power supply much larger than required, as long as the correct voltage is applied.
The reason for separate power supplies is to isolate the tube buffer from the voltage variation that will be caused by chip's demands on the power supply. The experts here can tell you all the bad stuff that can happen with the buffer and amplifier talking to each other via the power supply. The buffer is giving directions and doesn't want any argument or backtalk. As a neophyte, I'm guessing that there are ways to use the same supply transformer, but I bet the circuits necessary to give the buffer clean power from the amplifier chip supply would be more difficult to execute than just using a separate supply.
Sheldon
OK, separate trafos it is
Dave and Sheldon,
Thanks for that feedback. It looks like I'll go with separate trafos. I did note on Joe's site that some people have run these tubes at 24V and Joe recommends 35 so I thought 30V would be OK.
Since it appears I should use a completely separate power supply for the tube buffer and the current is low getting 35 (or even 75) volts to it won't be a problem.
I'm going to dig into the parts shelves today and see what I have and what I need to order.
Man, sometimes I wish I had never started doing this! Two tube amps, a few gainclones, three pairs of speakers all since January and I can't stop.
But I still need a tube pre-amp, this tube buffered gc and I haven't yet cut the MDF for my MLQWTs! Can't quit now!!!!
Dave and Sheldon,
Thanks for that feedback. It looks like I'll go with separate trafos. I did note on Joe's site that some people have run these tubes at 24V and Joe recommends 35 so I thought 30V would be OK.
Since it appears I should use a completely separate power supply for the tube buffer and the current is low getting 35 (or even 75) volts to it won't be a problem.
I'm going to dig into the parts shelves today and see what I have and what I need to order.
Man, sometimes I wish I had never started doing this! Two tube amps, a few gainclones, three pairs of speakers all since January and I can't stop.
But I still need a tube pre-amp, this tube buffered gc and I haven't yet cut the MDF for my MLQWTs! Can't quit now!!!!Re: Re: Single power supply
Hi Sheldon,
Nice analogy on V and I. Would using simple voltage regulators like LM317 and LM337 in the B+ for the tube stage help? The raw +-V could be taped off the supply of the gainclone. As the tube section is essentially low current, a post regulator capacitor would damp against the transients in the main supply.
Just some ideas, and maybe put a cap at the reference resistor for LM317 as a simple slow-start for the B+.
cheers
Sheldon said:
Hi Sheldon,
Nice analogy on V and I. Would using simple voltage regulators like LM317 and LM337 in the B+ for the tube stage help? The raw +-V could be taped off the supply of the gainclone. As the tube section is essentially low current, a post regulator capacitor would damp against the transients in the main supply.
Just some ideas, and maybe put a cap at the reference resistor for LM317 as a simple slow-start for the B+.
cheers
Re: Single power supply
Hi Guys
Sorry, been a bit out of the loop lately, a lot of water has passed under the bridge lately.
Yes, the isolation of a separate transformer is highly desirable for the reasons mentioned already, but the single largest reason is simpler than that.
The question seems to be repeating itself, can I avoid the second transformer? NO, you can't. Examine the design I've
posted carefully and see that the Tube Buffer is permanently ON, and that the power switch only activates the Gainclone chip. A delay is required to allow the filaments to stablise and the buffer to setle. Leave the buffer on at all times (it will sound better that way) and only switch GC On or Off.
Imagine if both Tube Buffer and GC coming ON together. Rather rude noises from you speakers. Not likely (I hope) to damage your speakers, but altogether unavoidable.
So keep that second transformer, it is not an expensive solution.
Joe R.
Sherman said:
Hi Guys
Sorry, been a bit out of the loop lately, a lot of water has passed under the bridge lately.
Yes, the isolation of a separate transformer is highly desirable for the reasons mentioned already, but the single largest reason is simpler than that.
The question seems to be repeating itself, can I avoid the second transformer? NO, you can't. Examine the design I've
posted carefully and see that the Tube Buffer is permanently ON, and that the power switch only activates the Gainclone chip. A delay is required to allow the filaments to stablise and the buffer to setle. Leave the buffer on at all times (it will sound better that way) and only switch GC On or Off.
Imagine if both Tube Buffer and GC coming ON together. Rather rude noises from you speakers. Not likely (I hope) to damage your speakers, but altogether unavoidable.
So keep that second transformer, it is not an expensive solution.
Joe R.
Re: Re: Single power supply
Hi Joe,
Thanks for the words of wisdom. Suppose if the startup delay is taken care, or a mute facility on the chip amp, or a separate small trans for the filament. Would taping off the same trans for B+ be an issue?
I really like your elegant design using +- 35v for the tube, my immediate reaction was the possibility using the same trans, maybe separate rectification circuits. It is less so a case of cost savings on the separate trans, but keeping only the necessary parts.
cheers
Joe Rasmussen said:
Hi Joe,
Thanks for the words of wisdom. Suppose if the startup delay is taken care, or a mute facility on the chip amp, or a separate small trans for the filament. Would taping off the same trans for B+ be an issue?
I really like your elegant design using +- 35v for the tube, my immediate reaction was the possibility using the same trans, maybe separate rectification circuits. It is less so a case of cost savings on the separate trans, but keeping only the necessary parts.
cheers
Re: Re: Re: Single power supply
There is always more than one way to skin a cat (what an awful thought), but you are right, my idea was to be as simple as possible, and elegant too? I'll take that as a complement. About using the mute function (which LM3875 doesn't have) it may be possible too. But my main argument here is that it's a good idea to leave the tubes permanently ON, it will sound better that way. Most audiophiles I know leave their preamps ON and only switch the power amps, this is pretty much the same thing.
One other thing: Some have suggested switching the DC to the Gainclone chip. But doing so would require a two pole switch. Both poles will NOT make contact at the precisely same time, they are not perfectly in sync, so for some miliseconds you will have a potential DC offset into speaker near full rail, either negative or positive, so only ever switch AC on the primary side to prevent that. All these consequences were considered in what looks simple, and is simple, but also reasonably safe.
Joe R.
Systeme_D said:
There is always more than one way to skin a cat (what an awful thought), but you are right, my idea was to be as simple as possible, and elegant too? I'll take that as a complement. About using the mute function (which LM3875 doesn't have) it may be possible too. But my main argument here is that it's a good idea to leave the tubes permanently ON, it will sound better that way. Most audiophiles I know leave their preamps ON and only switch the power amps, this is pretty much the same thing.
One other thing: Some have suggested switching the DC to the Gainclone chip. But doing so would require a two pole switch. Both poles will NOT make contact at the precisely same time, they are not perfectly in sync, so for some miliseconds you will have a potential DC offset into speaker near full rail, either negative or positive, so only ever switch AC on the primary side to prevent that. All these consequences were considered in what looks simple, and is simple, but also reasonably safe.
Joe R.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.