Upupa
As you seem to have a very good knowledge about the security norms:
What do you think about opening a separate thread about security? I would share, of course and I am interested, to know the norms.
Things like: how must I connect the ground to the box? Soldering, crimping?
Yesterday, I speaked with the local highend dealer about this topic. He told me, it should be best, to use an commercial, certified psu and not to exeed the voltages within the amp of 60VDC.
Maybe, thats one of the reasons, to isolate speaker bindings, when you have internal PSU > 60V eg +/-30VDC what could in the worst case appear at the output?
Franz
P.S.
I dont disregard security norms. One dead idiot is one to much!
As you seem to have a very good knowledge about the security norms:
What do you think about opening a separate thread about security? I would share, of course and I am interested, to know the norms.
Things like: how must I connect the ground to the box? Soldering, crimping?
Yesterday, I speaked with the local highend dealer about this topic. He told me, it should be best, to use an commercial, certified psu and not to exeed the voltages within the amp of 60VDC.
Maybe, thats one of the reasons, to isolate speaker bindings, when you have internal PSU > 60V eg +/-30VDC what could in the worst case appear at the output?
Franz
P.S.
I dont disregard security norms. One dead idiot is one to much!
Franz, this thematic circle is realy wide. I mean that everybody who can make something with electronic must have basic knowledges. Sometimes sure is not all quite clear, 'cos norms ( as I know ) are often written by rather " newspeak " 
and there is only a little examples for bigginers. Question is, how start this theme - I mean, that the best will be with some concrete examples ( questions ), but with this must come guys here, I don't know what they can . To your question : I will write at evening, 'cos now I must go away.
and there is only a little examples for bigginers. Question is, how start this theme - I mean, that the best will be with some concrete examples ( questions ), but with this must come guys here, I don't know what they can . To your question : I will write at evening, 'cos now I must go away.
You see, we did discuss safety aspects in the tube buffered thread...
BTW:
Yesterday, my most critical audio friend (actually subwoofer user, with this speakers http://www.audiodiskussion.de/foren/fullrange/msg.php?idx=14080 and here http://www.audiodiskussion.de/foren/fullrange/msg.php?idx=14081
listenend a first time to a valve buffered t-network inverted gainclone with FE103E in the Fostex recommended horn enclosure:
His very large smile told me all. He was specially impressed about the very good integrated bass and would very quick change all his equipment with this chain...
Now, I can go and sleep, after I told "you" this
BTW:
Dont miss this foto: http://www.audiodiskussion.de/foren/fullrange/msg.php?idx=14082
BTW:
Yesterday, my most critical audio friend (actually subwoofer user, with this speakers http://www.audiodiskussion.de/foren/fullrange/msg.php?idx=14080 and here http://www.audiodiskussion.de/foren/fullrange/msg.php?idx=14081
listenend a first time to a valve buffered t-network inverted gainclone with FE103E in the Fostex recommended horn enclosure:
His very large smile told me all. He was specially impressed about the very good integrated bass and would very quick change all his equipment with this chain...
Now, I can go and sleep, after I told "you" this
BTW:
Dont miss this foto: http://www.audiodiskussion.de/foren/fullrange/msg.php?idx=14082
Hi!
Sorry for resurrecting this thread, but I’d like to share my “Tube buffered, t-network GC” experience and concerns here.
Recently I have finished a GC according to Joe’s schematics, albeit with some minor changes.
So, the configuration is the following:
1. LM 3886
2. Tube-E88CC (one tube for both channels)
3. Tube power supply: +-40 v; CLC=6600uF (3x2200), 220 R, 2200uF, bypassed by 100nf on each rail (One 35 VA transformer with separate secondaries )
4. IC Supply- +-25, unregulated. Caps: 1000uf, 100uf, 100nf. (one 300 VA trafo with dual secondaries, 35 A bridge rectifiers-two per channel)
5. T-network: 3k9+18k, 10k/100R/10 k with 10 k form + IN. to Ground.
6. Point to point wiring.
First, some explanations:
1. I have used LM 3886 because I couldn’t obtain the 3875.
2. I used +-40v for the tube because I thought it might be better, since I am using E88CC .
3. I have tried bigger chip capacitors-even with Carlos snubber, didn’t like the sound. So I stuck to 1000 uf… With bigger caps, bass is fuller, but hi-midrange is muddied….
4. I thought it would be better to use bigger caps for tube supply, for lower ripple….
5. T-network? At first I have built a “regular” version with 1 meg. Resistors. Had almost zero offset, and then modified one channel with t-network. Liked it. More open, clearer sound. Offset: 10.8 mV.
6. Filament PSU: As described by Joe, no changes.
Sound? Great! Worth the trouble!
Concerns:
1. If mains power supply interrupts unexpectedly (it does happen), then tube filament will cool off pretty fast, while big tube PSU (2) caps may continue to apply some voltage on a cold tube. I am afraid that tube might be damaged. What can I do? I had an idea to put a relay which will connect, say, 10 ohm resistor across these caps when the power switches off, so these caps will be discharged fast…
2. Speaker protection: All speaker protection circuits that I have seen by now include a relay in a speaker path. I don’t want this. Is there some sort of protection which will disconnect the primary side of LM 3886 transformer, and won’t reconnect before it is reset manually? This sort of protection might be slower, but it could still be effective. Anyone knows such schematics?
3. I saw that some people have included a zobel. I did not. Wondering whether it is better to include it or not. Some folks (Peter Daniel, if I am correct) say that zobel harms the soundstage on his GC, so he did not include it. Some argue that it is definitely better and safer to mount a zobel. I am confused. What to do?
Minor complaint: I hear a very slight hum when I place my ear next to the speaker. It might be grounding issue…
Otherwise, it’s gorgeous! It’s still not finished; everything is on my desk, waiting for a suitable home…
Best regards,
Vix
Sorry for resurrecting this thread, but I’d like to share my “Tube buffered, t-network GC” experience and concerns here.
Recently I have finished a GC according to Joe’s schematics, albeit with some minor changes.
So, the configuration is the following:
1. LM 3886
2. Tube-E88CC (one tube for both channels)
3. Tube power supply: +-40 v; CLC=6600uF (3x2200), 220 R, 2200uF, bypassed by 100nf on each rail (One 35 VA transformer with separate secondaries )
4. IC Supply- +-25, unregulated. Caps: 1000uf, 100uf, 100nf. (one 300 VA trafo with dual secondaries, 35 A bridge rectifiers-two per channel)
5. T-network: 3k9+18k, 10k/100R/10 k with 10 k form + IN. to Ground.
6. Point to point wiring.
First, some explanations:
1. I have used LM 3886 because I couldn’t obtain the 3875.
2. I used +-40v for the tube because I thought it might be better, since I am using E88CC .
3. I have tried bigger chip capacitors-even with Carlos snubber, didn’t like the sound. So I stuck to 1000 uf… With bigger caps, bass is fuller, but hi-midrange is muddied….
4. I thought it would be better to use bigger caps for tube supply, for lower ripple….
5. T-network? At first I have built a “regular” version with 1 meg. Resistors. Had almost zero offset, and then modified one channel with t-network. Liked it. More open, clearer sound. Offset: 10.8 mV.
6. Filament PSU: As described by Joe, no changes.
Sound? Great! Worth the trouble!
Concerns:
1. If mains power supply interrupts unexpectedly (it does happen), then tube filament will cool off pretty fast, while big tube PSU (2) caps may continue to apply some voltage on a cold tube. I am afraid that tube might be damaged. What can I do? I had an idea to put a relay which will connect, say, 10 ohm resistor across these caps when the power switches off, so these caps will be discharged fast…
2. Speaker protection: All speaker protection circuits that I have seen by now include a relay in a speaker path. I don’t want this. Is there some sort of protection which will disconnect the primary side of LM 3886 transformer, and won’t reconnect before it is reset manually? This sort of protection might be slower, but it could still be effective. Anyone knows such schematics?
3. I saw that some people have included a zobel. I did not. Wondering whether it is better to include it or not. Some folks (Peter Daniel, if I am correct) say that zobel harms the soundstage on his GC, so he did not include it. Some argue that it is definitely better and safer to mount a zobel. I am confused. What to do?
Minor complaint: I hear a very slight hum when I place my ear next to the speaker. It might be grounding issue…
Otherwise, it’s gorgeous! It’s still not finished; everything is on my desk, waiting for a suitable home…
Best regards,
Vix
Hi Vix
Great, your report! Thanks.
You should try about +/- 60V for the E88CC. Its about the best point in my experience.
Just use a voltage doubler...
Franz
P.S.
Are you dutch? In Amsterdam, there is one of my tube buffered t-net gainclones playing... I could send you the addresse, to compare the amps
Great, your report! Thanks.
You should try about +/- 60V for the E88CC. Its about the best point in my experience.
Just use a voltage doubler...
Franz
P.S.
Are you dutch? In Amsterdam, there is one of my tube buffered t-net gainclones playing... I could send you the addresse, to compare the amps
I use a delay circuit with a NE555. While powering up (20 sec) or down (immediately), the input of the LM3... is grounded by relay.
NE555 circuit
I hope, this helps
Franz
--------------------------------------------
Inverted GC's: never without t-net...
Hi Franz!
Thanks for nice suggestions. Yes, I made a mistake, I meant CRC not CLC (no inductor there). As for the tube, there are two options: Either to increase the voltage to +-60 (this won’t be easy since my caps are 50V ones). Or, there is a possibility to obtain the 6922 tube and operate it at existing supply, +-40 v.
Speaker protection is still puzzling me. I can and I will use an NE 555 for a delay to avoid turn on thump, but the idea was to build a protection that will save the speakers should the LM 3886 fail. But I’d rather like to do this without including a relay in the speaker path.
Two things that I still need to do to eliminate that hum: Ground pin 9 (I did not ground it yet) and I did not ground the filament circuit. In fact, I use a 7805 regulator, but I increase its voltage by adding 3 1n4148 diodes in series with 7805 ground pin, + 3 ohm resistor in series with the output….As a final result I get 6.2 v regulated DC, so no problem here.
Franz, did you use a zobel on the output? I am still wondering whether to include it or not.
Did you use a regulated PSU for LM 3886 (3875)? Did you try higher capacitance and Carlos snubber? I was thinking about regulated PSU with LM 338, but I might have too low voltage after regulation…
Thanks for a support!
Vix
Thanks for nice suggestions. Yes, I made a mistake, I meant CRC not CLC (no inductor there). As for the tube, there are two options: Either to increase the voltage to +-60 (this won’t be easy since my caps are 50V ones). Or, there is a possibility to obtain the 6922 tube and operate it at existing supply, +-40 v.
Speaker protection is still puzzling me. I can and I will use an NE 555 for a delay to avoid turn on thump, but the idea was to build a protection that will save the speakers should the LM 3886 fail. But I’d rather like to do this without including a relay in the speaker path.
Two things that I still need to do to eliminate that hum: Ground pin 9 (I did not ground it yet) and I did not ground the filament circuit. In fact, I use a 7805 regulator, but I increase its voltage by adding 3 1n4148 diodes in series with 7805 ground pin, + 3 ohm resistor in series with the output….As a final result I get 6.2 v regulated DC, so no problem here.
Franz, did you use a zobel on the output? I am still wondering whether to include it or not.
Did you use a regulated PSU for LM 3886 (3875)? Did you try higher capacitance and Carlos snubber? I was thinking about regulated PSU with LM 338, but I might have too low voltage after regulation…
Thanks for a support!
Vix
I used a RL network like suggested from nationals datasheets for commercial amps. Just to avoid some complaints.
But not for my speaker setup. I dont need something (short speaker cables, fullrange speakers, no passive elements).
Just test it!
I have both variants. With the unregulated variant, it is essential to use a tranny with >=225VA, at least 2x10000uF AND snubbers.
I am sure: try to work with different working points for the tube, as it could be very "sound shaping". Even with the 6922 the voltage / current is too low, as I can assure you. And many 6922 tend to have strong microphony.
Maybe, it is worth to use 63V or 100V caps for the tube psu?
Why not a regulated and adjustable psu for the tube, to find the optimal working point by listening?
Franz
Vix said:
My question, along similar lines, did you use the 0.22R resistor on the output?
I too share extreme reluctance to put a series relay on the output. I can imagine this verbal exchange:
Hey, I like the sound of your speaker cable. Back come the answer: Hey, it's the sound of my super-duper relay.
I think I prefer listening to speaker cables rather than relays.
Joe R.
Hi
Joe, I use a two 0,1 5w resistors in series on the output ( I just happend to have these values). So I'd rather not include a zobel.
Relay? If it's brand new, maybe there will be no problems.
But imagine what happnes to the sound when relay contacts wear out....
I had a chance to listen to a speaker with poor connections ...AWFUL.
On the other hand, the scene of burning speakers is even more horrible...
Joe, I use a two 0,1 5w resistors in series on the output ( I just happend to have these values). So I'd rather not include a zobel.
Relay? If it's brand new, maybe there will be no problems.
But imagine what happnes to the sound when relay contacts wear out....
I had a chance to listen to a speaker with poor connections ...AWFUL.
On the other hand, the scene of burning speakers is even more horrible...
Vix said:
May I suggest ditching the resistors. Get some 10 Ohm to 22 Ohm 1 Watt resistors, get some enamelled copper wire about 0.25mm and wind them some 15-20 turns (or as close as you can get to that) around the 1 Watt resistors. The enamelled wire is either stripped or use a very hot iron to melt away enamel (it will smell) and solder this coil to each end of the 1 Watt resistors. Repeat for both channels. I then use a bit of heatshrink to keep the wire in place. Or use some insulation tape wrapped a few times around the resistors.
Let me know what you think of it after you've had a listen. Should be no need for a zobel.
Joe R.
PS: You can even use wire wrap wire for this. Don't remove the insulation except on the ends to soder. In this case, put on as many turns as you can.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.