Thanks Joe
I will try that tonight. So in this case 0,22 ohm resistor will not be necessary, right?
The only unresolved issue is speaker protection. Do you use it? It seems to me that you don't include it in your amps. We don't like the "sound of rusty relay contacts"🙂
My idea was to find a circuit which, once significant DC on the speaker output is detected, turns off the main transformer and does not re-connect until it is done manually. (and after the cause of error is removed😀 .
But I haven't found such a thing yet....
Vix
I will try that tonight. So in this case 0,22 ohm resistor will not be necessary, right?
The only unresolved issue is speaker protection. Do you use it? It seems to me that you don't include it in your amps. We don't like the "sound of rusty relay contacts"🙂
My idea was to find a circuit which, once significant DC on the speaker output is detected, turns off the main transformer and does not re-connect until it is done manually. (and after the cause of error is removed😀 .
But I haven't found such a thing yet....
Vix
Vix said:
That's right. use in place of current resistors.
Re use of relay, no I don't use them. But if I was to make a suggestion, perhaps it would be this. With 1000uF ps caps, the stored energy is lower than normal. Find a circuit that monitors DC on the output, that activates the AC side of the transformer. So if a fault condition occurs (potential DC into the speakers), the AC side of the primary is shut down and only the limited energy stored in 1000uF caps is the remaining danger to the speakers, and being smaller than usual this is an advantage.
DO NOT use twin relay on the DC side, to shut down the amp. This will cause a DC spike into the speakers as the two contacts will never be 100% synced.
Joe R.
To Joe : Try to short uploaded 1000 M cap by speaker terminals - you will see ( and hear ) how low is energy acumulated in this low capacity 😀 .
Hi
Joe, exactly as I thought! Relay should disconnect the primary AC side of the transformer.
Pavel, I think that even your concern about 1000uF caps can be resolved.
I would mount four more relays ( or one with more separate contacts), so, when activated, each of them would connect a 10 ohm resistor across each power supply cap, which would discharge them.
As a result, even if some energy leaks into speakers, it shouldn't be able to harm them😀
Vix
Joe, exactly as I thought! Relay should disconnect the primary AC side of the transformer.
Pavel, I think that even your concern about 1000uF caps can be resolved.
I would mount four more relays ( or one with more separate contacts), so, when activated, each of them would connect a 10 ohm resistor across each power supply cap, which would discharge them.
As a result, even if some energy leaks into speakers, it shouldn't be able to harm them😀
Vix
To Vix : When is relay activated ? If is DC on output. When id DC on output ? If is broken some output device ( in mostly cases ). After which time is relay activated and what is the switch off time of this relay ? Cca 50 - 100 msec or more. What make by this time PS ? Is shorted by bass speaker. How is shorting current of 300 VA transformer, if is shorted by 3 Ohm winding of speaker ? Cca 10 A. What cause 10 A of DC current by the time 100 msec with speaker choke ? Burn it. Etc.... and it is not all problems by this solution. But if you can, make it. 😎
Pavel,
If you are concerned about 50-100 ms switch-off time, I think that you might have the same problem even if the relay is in the speaker path.
Dc sensing circuit would be similar to the ones that use relay in the speaker path, except the fact that it will disconnect the AC primary of the main transformer. DC sensing circuit would be powered by a separate transformer. The only difference would be that the "new" circuit would not re-connect automatically, but would keep the circuit disconnected until manual reset.
Vix
If you are concerned about 50-100 ms switch-off time, I think that you might have the same problem even if the relay is in the speaker path.
Dc sensing circuit would be similar to the ones that use relay in the speaker path, except the fact that it will disconnect the AC primary of the main transformer. DC sensing circuit would be powered by a separate transformer. The only difference would be that the "new" circuit would not re-connect automatically, but would keep the circuit disconnected until manual reset.
Vix
To Vix : Basicaly you have thruth, but in your case is all charge of supply discharged through the reprobox after switch off of primare, in second case not. Relay's switch off time is shorter.
IGC and active crossover biamped design
Hello to all,
I have read through first 2/3ds of this thread last night and all of T-network and got (needless to say) very inspired.
I've made small 2-way active speakers with 24dB/oct active filters (3xOPA2134) using two pairs of standard TDA7294, each pair of amps inside the loudspeaker with heatsink on the back panel. To be honest, I naturally dropped a tear during first listening as the music finally just came out of the boxes to hug me.
However, when judging more critically despite being quite detailed TDA7294 lacks some of "natural ease". I'd say they sound a bit "sturdy".
So I'm going to build a pair of IGCs (without input buffer due to existing crossovers) with T-network and then make a comparison of amps, and do the same with slightly different set of speakers after finishing my first T-Lines.
What you would suggest in biamped scenario: one LM4780 per channel or two LM3886/3875s with separate rectifiers/filter caps? I'm weighting more complexity of two chips per channel against possible positive impact of current draw separation for high frequency channel. I doubt that I'll be using them at more than 15-25 Watts for 95% of time.
Prtg
Hello to all,
I have read through first 2/3ds of this thread last night and all of T-network and got (needless to say) very inspired.
I've made small 2-way active speakers with 24dB/oct active filters (3xOPA2134) using two pairs of standard TDA7294, each pair of amps inside the loudspeaker with heatsink on the back panel. To be honest, I naturally dropped a tear during first listening as the music finally just came out of the boxes to hug me.
However, when judging more critically despite being quite detailed TDA7294 lacks some of "natural ease". I'd say they sound a bit "sturdy".
So I'm going to build a pair of IGCs (without input buffer due to existing crossovers) with T-network and then make a comparison of amps, and do the same with slightly different set of speakers after finishing my first T-Lines.
What you would suggest in biamped scenario: one LM4780 per channel or two LM3886/3875s with separate rectifiers/filter caps? I'm weighting more complexity of two chips per channel against possible positive impact of current draw separation for high frequency channel. I doubt that I'll be using them at more than 15-25 Watts for 95% of time.
Prtg
Hi,
I have built an amp using TDA 7294 two years ago. I didn't like it and still don't like it. It sounds...COLD. As if musicians went to the other room. I felt dizzy after 10 mins of listening. Give lm 3886 a chance. It might surprise you.
Vix
I have built an amp using TDA 7294 two years ago. I didn't like it and still don't like it. It sounds...COLD. As if musicians went to the other room. I felt dizzy after 10 mins of listening. Give lm 3886 a chance. It might surprise you.
Vix
Thanks for encouragement, taming the 3886 right now 😎
I'm sorry if I this was already asked, but have anyone made measurements to prove if low-pass filtering at input stage is still needed with 3886/4780? Do they show the same ringing behaviour at the same frequency (over 320kHz) as noticed by Joe with 3875?
I'm sorry if I this was already asked, but have anyone made measurements to prove if low-pass filtering at input stage is still needed with 3886/4780? Do they show the same ringing behaviour at the same frequency (over 320kHz) as noticed by Joe with 3875?
PRTG said:
Don't worry about the measurements, much better is to try it, on the basis of what it does sound wise. I have often repeated that it should be tweaked by ear and not measurement. Keep the resistor values the same and try a few different value caps.
Try it and you may well like what it does. Others have.
Joe R.
Hello,
Here is a shematic from an LM4780 with vaccum tube. 😉
I have write it while thinking of the homepage of our newest member - Joe Rasmussen on "the JLTi Tube Hybrid Amplifier".
I would like that you say to me that do you think about my shéma.
(Oh... sorry for my poor english)
A lot of thanks
Almu
Here is a shematic from an LM4780 with vaccum tube. 😉
I have write it while thinking of the homepage of our newest member - Joe Rasmussen on "the JLTi Tube Hybrid Amplifier".
I would like that you say to me that do you think about my shéma.
(Oh... sorry for my poor english)
A lot of thanks
Almu
Attachments
By this working point of tube have it relatively big distortion. Ua is low and low is also input impedance of power amp. And gain of amp is redundant - try to make it lower, change 18 k up to cca 47 k.
If I'm not mistaken, that's essentially the same tube buffer that Joe first published right? They sound very good and a good match in taming a chipamp.
If something " sound good ", it must not be " end station ", sometimes is improvement possible. But if you are satisfied, let it be 😎 . BTW in many times was connections fifty ears ago better than today's, 'cos at this time constructeurs had know, what they make 😀 .
Hello all,
Last night I've assembled one channel of 3886 based IGC. Assembly of second one today took less than an hour in a rush. Oddly enough, both worked without major flaws from first switch-on, therefore my test speaker (for all my blow-out needs) stayed intact 😎 I hope to make more serious testing during next couple of weeks.
The history: I actually re-used the ICs and PCBs I made for standart 3886 about 4 years ago. The goal was to bridge them, but I didn't get so far. After 2 out of 4 were self-destructed without appparent reason. I stopped testing and switched to TDAs. I guess I used something like +-40 Volts back then which could be too much for amp to remain stable at given gain. Rewiring was easy as unbuffered IGC actually has less elements than National recommended schematics.
And, yes, tomorrow is my birthday, and this pair of tweaked GCs is nice gift I gave to myself, thanks to all of you. So, have your preffered drink 🙄
Prtg
P.S. Joe, some of my relatives live in Australia. I'd like to go and meet them finally someday. Hopefully next year. There are also others in USA and Siberia, too, so I must decide.
Last night I've assembled one channel of 3886 based IGC. Assembly of second one today took less than an hour in a rush. Oddly enough, both worked without major flaws from first switch-on, therefore my test speaker (for all my blow-out needs) stayed intact 😎 I hope to make more serious testing during next couple of weeks.
The history: I actually re-used the ICs and PCBs I made for standart 3886 about 4 years ago. The goal was to bridge them, but I didn't get so far. After 2 out of 4 were self-destructed without appparent reason. I stopped testing and switched to TDAs. I guess I used something like +-40 Volts back then which could be too much for amp to remain stable at given gain. Rewiring was easy as unbuffered IGC actually has less elements than National recommended schematics.
And, yes, tomorrow is my birthday, and this pair of tweaked GCs is nice gift I gave to myself, thanks to all of you. So, have your preffered drink 🙄
Prtg
P.S. Joe, some of my relatives live in Australia. I'd like to go and meet them finally someday. Hopefully next year. There are also others in USA and Siberia, too, so I must decide.
PRTG said:
I love to have visitors. I live in the South-West of Sydney. Even in winter time it is usually sunny and often warms up to 20C during the day. Great city, great harbour and great opera house. It's cheaper than most western cities. But beware, we drive on the wrong side. 😎
Joe R.
Gentlemans,
Could you be so kind to advise me how to use/replace LM 3875 with LM 3886 in Mr. Joe Rasmussen IGC tube hybrid schematic ?
This is the only chip available in this part of the world ,unfortunattely .
After reading all your excelent posts ( yeah I become a fan of this forum ) I'm still not sure what should I have to change in order to fit 3886 instead 3875
Mr.Vix wrote something about using 3886 and ECC88 or 83 which is my case too. Can I hope for a helping hand ?
Sorry for my bad english
Thank you,
Best Regards
Could you be so kind to advise me how to use/replace LM 3875 with LM 3886 in Mr. Joe Rasmussen IGC tube hybrid schematic ?
This is the only chip available in this part of the world ,unfortunattely .
After reading all your excelent posts ( yeah I become a fan of this forum ) I'm still not sure what should I have to change in order to fit 3886 instead 3875
Mr.Vix wrote something about using 3886 and ECC88 or 83 which is my case too. Can I hope for a helping hand ?
Sorry for my bad english
Thank you,
Best Regards
Once you deal with the mute the 3886 should connect pretty much the same as 3875... the ECC88 = 6922 = 6DJ8 = 6H23pi
dave
dave
- Status
- Not open for further replies.