Tube with Power IC Output Stage - JLTi

[l_paitoon]

Tube IGC (Inverted Gainclone) - Need help

I have followed and built the Tube Buffer circuit based on the JLTi DIY version to HUMMMMMM!!!!!!!. My PSU filament use LT1085, PSU +-35V for 6922 using 470uf on the front,R 220 series and 68000uf on the final, with 1uf film caps for bypass.

 

[sandro600]

I have followed and built the Tube Buffer circuit based on the JLTi DIY version to HUMMMMMM!!!!!!!.

do you mean you have hum ? And do you really have 68000uf on the final ?

Ive just completed a single channel prototype of the JR JLTi .....
on first firing it up it had hum ++.
But I managed to get rid of it almost completely by moving the ground lead from the 35V tube power supply so that it connected with the G ground closer to the negative terminal of the speaker. Before the change it was connected to the G closer to the power supply. So ,like others have suugested, by playing with the grounding lead position, you may get rid of it.

Joe has also suggested adding capacitance to the tube filament supply (the 6.3V suply) to a total of 2200uf.

Its worth persevering > the sound is very rich. Bass reproduction in particular is very detailed. Its all very encouraging .
Now for the second channel.

 

[sandro600]

How is the sound altered by changing the value of the 1n3 cap ?

At the moment my prototype channel has 2x680pf in parallel
The bass is very detailed and rich. The upper frequencies not as prominent.

If I put a lower value in does this affect the higher frequencies ?

 

[steve]

Hi..

Has anyone attempted Joe's high powered version posted on page eleven of this thread?

I have speakers that are pretty difficult to drive...( My fifty watt Creek (4240se) runs out of breath before my speakers do (speakers are DIY MTM's with scanspeak 8545's and 9300)).

So I'm trying to decide between going with a high powered version (Joe's bridged-parallel) or multi-amping with an active crossover (Pedjas regulator, Joes tube buffer, Moes MOX, Brians 3875 modules (3 per channel).

Any opinions as to which route is likely to provide best sound quality would be greatly appreciated..

Thanks
Steve

 

[leadbelly]

Any opinions as to which route is likely to provide best sound quality would be greatly appreciated..

Bi- or tri-amping is much more likely to provide best SQ.

And if you go and build JR's big amp, I would not build it in 1 case. Those are some pricey components, and you would likely be happier making it in modular form so that you can play around with your audio chain without major surgery.

 

[Joe Rasmussen]

Hi..

Has anyone attempted Joe's high powered version posted on page eleven of this thread?

I have speakers that are pretty difficult to drive...( My fifty watt Creek (4240se) runs out of breath before my speakers do (speakers are DIY MTM's with scanspeak 8545's and 9300)).

So I'm trying to decide between going with a high powered version (Joe's bridged-parallel) or multi-amping with an active crossover (Pedjas regulator, Joes tube buffer, Moes MOX, Brians 3875 modules (3 per channel).

Any opinions as to which route is likely to provide best sound quality would be greatly appreciated..

Thanks
Steve

I just want to point out that this was posted as a thought experiment, although at the time I was hoping a friend was going to let me build it for him. Alas, that is yet to happen.

Joe R.

 

[Fishy]

Hello ...

I am using a xfmr with 120VCT secondary for the tube (the plate/cathode). So I will end up with ~ +/-85VDC .... so, do I need to change the 10K/1W cathode resistor (1) to a higher wattage (2) to a higher resistance? What should I consider here given the much bigger voltage I am applying?

Thanks!!

Horace

 

[Joe Rasmussen]

Hello ...

I am using a xfmr with 120VCT secondary for the tube (the plate/cathode). So I will end up with ~ +/-85VDC .... so, do I need to change the 10K/1W cathode resistor (1) to a higher wattage (2) to a higher resistance? What should I consider here given the much bigger voltage I am applying?

Thanks!!

Horace

Simple Ohm's Law says: If original circuit the current was 3.7mA approx then...

85/0.0037 = 22972 Ohm

So increasing 10K to 22K is about right. What about dissipation (Wattage)?

85 times 0.0037 = 0.3145 Watt

Since I believe that any resistor should not exceed 40% of its rating in tube gear (be conservative due to higher ambient temps) then 0.3145 Watt is still within that (31.45% of 1W). So go ahead and use 22K 1W.

Joe R.

 

[Fishy]

Simple Ohm's Law says: If original circuit the current was 3.7mA approx then...
85/0.0037 = 22972 Ohm

Thanks ...

I was sure the one thing I was missing when I first asked the question is the expected current. Was ~3.7mA a chosen value or it's a characteristic of the tube? (I was more inclined to think the value is "by design") So, the aim is to make the cathode at zero volt ... and what aspect of the circuit design makes sure that the grid is at a lower voltage than the cathode?

Sorry if those questions are tooo fundamental. 🙄 even a simple circuit like the cathode follower is kinda new to me

 

[Joe Rasmussen]

Was ~3.7mA a chosen value or it's a characteristic of the tube? (I was more inclined to think the value is "by design")

Originally with 35-36V rails plus 1 volt for cathode bias voltage (which point it finds automatically) meant about 37V across 10K. I could have increased current as some might have deemed better, but I didn't want to use anything lower than 10K on the cathode, so it ended up being 3.7mA and also the output impedance with the 6922 I was using, a quite low 200 Ohm output Z.

So, the aim is to make the cathode at zero volt ... and what aspect of the circuit design makes sure that the grid is at a lower voltage than the cathode?

The cathode voltage (its bias point) is about 1 volt, so if you wanted to be absolutely exact, you would use 86V rather 85V, to calculate current across cathode value, but needless to say, not that critical.

If you want to, get the circuit running with 22K 1W. Then try upping the current by using 15K 2W (5.5mA) and 10K 2W (8.5mA) and report back if you felt there was any improvement. That is what DIY is all about.

Joe R.

 

[lumanauw]

Hi, Mr. Rasmunssen,

Wanted to ask you this. How much lower the tube can go? I'm planning to built preamp version of your tubed design. The preamp is with opamp, so the supply is only +/-15V. Will the tube works properly with +/-15V, or I will have to make other supply just for the tube?

 

[Rarkov]

Hi,
Is it possible to use the IGC part of the JLTi circuit without the tube? I guess that I just remove the tube section upto but not including the 3u3 Capacitor?

Thanks for any help...
Gaz

 

[Joe Rasmussen]

Will the tube works properly with +/-15V, or I will have to make other supply just for the tube?

I have heard of the 6922 working as low as 24V, but I wouldn't recommend it. As for 15V, I doubt it very much.

Joe R.

 

[Joe Rasmussen]

Hi,
Is it possible to use the IGC part of the JLTi circuit without the tube? I guess that I just remove the tube section upto but not including the 3u3 Capacitor?

Thanks for any help...
Gaz

The answer is maybe.

What happens if accidentally you turn it on and there is no input or the interconnect/RCA connection goes wonky? In that case the input resistor (and cap) is un-connected. What then? The IGC reverts to UG (Unity Gain) and since it isn't UGS (Unity Gain Stable) it will likely go very unstable right into your very vulnerable speakers giving them a traumatic experience.

Now maybe, and only maybe the 1n3 shown in the circuit may prevent that by keeping the 'gain' above 20 dB (the point below which NS does not guarantee stability) at high frequencies, but I haven't tested it.

As for the cap, yes it must stay, unless you like DC into your speakers.

You may like to look at this page, sans tube variations:

http://members.ozemail.com.au/~lisaras/sanstube.htm

Joe R.

PS: "sans" means 'without.'

 

[Rarkov]

Hi Joe,
Thanks for the reply. I had read the whole of your site when I was building my Tube IGC, but I saw no mention on the Sans Tube page about keeping the circuit stable.


I am building monoblocks without volume control, so I have two questions:

1) Assuming I take off the tube section upto the 3u3 cap (but not including it - in otherwords, it will stay in the circuit), how do I keep the circuit stable? A 22k to ground at the input - or something similar?

2) What should the capacitor value be after the 2k2 (IIRC) if there is not volume pot present?

Thanks again,
Gaz

 

[Nuuk]

Rarkov, Joe designed that IGC circuit to go with the valve buffer. Why on earth do you want to use it without when there are countless other GC circuits that work well without the buffer?

Or do you want to include the LPF but not build a tube buffer? In that case, it isn't much work to add a simple opamp buffer instead! 😉

 

[Peter Daniel]

A 22k to ground at the input - or something similar?

As long as you have input shunted by some sort of resistance (in the above range), you shouldn't experience any major problems if nothing is connected to the amp.

 

[Rarkov]

Thanks Peter,
I'll do that then! 😉 I'm supposed to be writing my dissertation on "A method of measuring frequency response of audio circuits" - but I can't think straight, so I am building an amp instead! 🙂

Nuuk,
I built a buffered IGC - but didn't like it very much, so I am building a tiny (!) monobloc using the P2P wired LM3875 part of the TB IGC. I don't like the maintenance side of tubes. I like to leave my amps on all day long - as they should be really.

For my TB IGCs, I used a PIC microprocessor to sense when there was an input into the amp and switch on the amps accordingly. It would also sort out the switch on and turn off of the amps for no noise. If there was no input for 10 minutes, it'd switch off, and if a signal was applied after it had gone into standby, it would start the 15 second cold start again...

PIC control is the only way tubes and I will get on I'm afraid...This is no reflection on the sound however, which was very nice - on the one channel I constructed before giving up on the boring P2P wiring on strip board! I will probably try again with a circuit board since I have spent £50 on the cases.

I only ever seem to make monoblocs! I should really learn - monoblocks = 2 x stereo!

Thanks,
Gaz

 

[Peter Daniel]

Why don't you build the non inverted monoblocks then? Many people who tried the non inverted configuaration, actually prefer it to inverted one. If you don't want to bother with buffer stage, the NIGC is actually recommended, as it seems to be working better with directly connected sources.

 

[Rarkov]

This is not on my "project list" and is taking up as little time & money as possible. I had already built the IGC modules for the TB IGC so I wanted to recycle them. I'd already used Audio Note resistors at £1 per piece and didn't want them going to waste! 🙂

Gaz