Upupa,
Who makes inverting amps inverts the speaker wires inside the amp.
So, the output is in phase.
At least I've always done that.
You're right about the speakers sucking air instead af spitting it at a stroke of a drum.😀
Who makes inverting amps inverts the speaker wires inside the amp.
So, the output is in phase.
At least I've always done that.
You're right about the speakers sucking air instead af spitting it at a stroke of a drum.😀
Peter Daniel said:
Upupa Epops said:
And what about the speakers themselves? It may be simple with full rangers and two-way 1st order XO systems but it stops to be that simple otherwise. Anyone bothers with this?
By "in phase" I actually meant simple 2 way with 1st order filter and in phase tweeter.
With 3 way, it's more complicated. Usually midrange is in phase and sub/tweeter according to what measures/sounds better.
With 3 way, it's more complicated. Usually midrange is in phase and sub/tweeter according to what measures/sounds better.
I do not know if there is any standard about this. Midrange’s polarity may be swapped instead. The things change if midrange and tweeter are or are not corrected for the acoustical offset, 2-3cm forward or back may make that difference in the way they are better if in or out of absolute phase. The only constant I have found for now is that a bass and a midrange should be connected the opposite way, and then what should have swapped leads? Yet this is only my experience and this is only assuming still 1st order XO for all drivers.Peter Daniel said:
I’d say that, before the “my DAC/amp reverse the absolute phase so I swapped the leads on the speaker to ensure correct phase” step, one should know the situation with own speaker.
Pedja
Franz, I ran the amp for two hours continuously and then re-checked the offset voltages. They were 5.2 mV and 1.2 mV.
This is also the coolest running GC that I have built. Is that something to do with the T-network? What have you other T-networkers found as regards chip temperature?
Mines all sealed up and locked away so I couldn't tell you how hot it runs.
Got my latest set of samples today, so hopefully will be able to compare regulation to simple power supplies. Also this one will have OPA627 buffers against my current amp which has valve buffers.
Shoog
Got my latest set of samples today, so hopefully will be able to compare regulation to simple power supplies. Also this one will have OPA627 buffers against my current amp which has valve buffers.
Shoog
Nuuk said:
Oops, sorry about the omission. Rkomp (per Franz's spreadsheet) is 10.0k. Gain comes to 44. All resistors 1% and checked for accuracy.
Sheldon
Not to bad, this spreadsheet... 😉
A very good result!
So, in fact, we really could try to minimize bias by trial. The thermal effects seems to be in a aceptable range.
Franz
I apologize:
I just reversed the speaker connections (Fostex FE103E in recommended enclosure):
There is a remarkable difference, in the way to connect them.
Sorry, Peter!
I just reversed the speaker connections (Fostex FE103E in recommended enclosure):
There is a remarkable difference, in the way to connect them.
Sorry, Peter!
As long as we come to right conclusions (and improvement in our systems) I don't see a problem with any discussion 😉
Re: Another T
Correction: The above was with the mute on. Under operating conditions, the values are 40mv and 12mv. As noted previously, the return of the non-inverting input to ground was via 10.0K.
Sheldon
By the way, the 3876 chip has a ground pin on the chip. Any guess as to the proper site for this ground. I've got it connected to the signal ground. I've got a little hum, but have not yet done any sorting.
Sheldon said:
Correction: The above was with the mute on. Under operating conditions, the values are 40mv and 12mv. As noted previously, the return of the non-inverting input to ground was via 10.0K.
Sheldon
By the way, the 3876 chip has a ground pin on the chip. Any guess as to the proper site for this ground. I've got it connected to the signal ground. I've got a little hum, but have not yet done any sorting.
Re: Re: Another T
Recorrection: sorted out some joint/ground issues (thanks Carlos, the data sheet wasn't clear there). Now have 10mv DC on one channel, 12.5 on the other. About 3.5mv AC on one side, 5mv on the other (60/120 hum), so still some minor ground checking. I quickly threw together the test assembly (I'm waiting for parts to do a remote), so I think I can do a little better with idle noise.
Sheldon
Sheldon said:
Recorrection: sorted out some joint/ground issues (thanks Carlos, the data sheet wasn't clear there). Now have 10mv DC on one channel, 12.5 on the other. About 3.5mv AC on one side, 5mv on the other (60/120 hum), so still some minor ground checking. I quickly threw together the test assembly (I'm waiting for parts to do a remote), so I think I can do a little better with idle noise.
Sheldon
Re: Re: Re: Another T
Join the two channels' power star ground with a short and thick wire.
From the middle point of that wire connect the ground wire from the PSU.
Sheldon said:
Join the two channels' power star ground with a short and thick wire.
From the middle point of that wire connect the ground wire from the PSU.
Re: Re: Re: Re: Another T
Thanks Carlos,
Did a bit of tidying up. Now less than 2mv AC each channel. I think that's about as good as I can do with these boards (from Brian GT's original offer) and my layout. But it's pretty good, as no audible noise until within a foot of the speaker. Amp sounds very fine, but no careful comparative tests yet. BTW this amp has Pedja's reg. (last discreet version) and buffer. Not using the mute pin, as I made separate reg. for the buffer (same as the other but less capacitance and minus the final power transistor), powered by a separate transformer (Hammond 164D28) and small bridges. No power on thump, as the buffer reg. is connected direct to the mains and the power reg. is switched, ala Joe's buffer set up.
Later, I may try a speaker protection routine via the mute pin with a dual op-amp (tl 082, some diodes, resistors, and a NE555). This could be ugly, cause I'm winging it here.
By the way, Franz's set up with the NE555 still requires a solenoid, so not much simpler than my solenoid approach. Though it does offer easier control over a wide range of times. I think the 555 probably can be rigged to pull the mute pin to ground directly, but I haven't looked it over yet.
Sheldon
carlosfm said:
Thanks Carlos,
Did a bit of tidying up. Now less than 2mv AC each channel. I think that's about as good as I can do with these boards (from Brian GT's original offer) and my layout. But it's pretty good, as no audible noise until within a foot of the speaker. Amp sounds very fine, but no careful comparative tests yet. BTW this amp has Pedja's reg. (last discreet version) and buffer. Not using the mute pin, as I made separate reg. for the buffer (same as the other but less capacitance and minus the final power transistor), powered by a separate transformer (Hammond 164D28) and small bridges. No power on thump, as the buffer reg. is connected direct to the mains and the power reg. is switched, ala Joe's buffer set up.
Later, I may try a speaker protection routine via the mute pin with a dual op-amp (tl 082, some diodes, resistors, and a NE555). This could be ugly, cause I'm winging it here.
By the way, Franz's set up with the NE555 still requires a solenoid, so not much simpler than my solenoid approach. Though it does offer easier control over a wide range of times. I think the 555 probably can be rigged to pull the mute pin to ground directly, but I haven't looked it over yet.
Sheldon
Re: Re: Re: Re: Re: Another T
You mean you are using a board made for the LM3875 with the LM3886?
Anyway, always glad to help.
Sheldon, with single PSU feeding two channels you need to get them as close as possible and make that thick wire joining the two channel's ground as short as possible. It's really a question of centimeters.
I don't have any noise with my LM3886 power amp (the mini-Krell😀 ), but I think that for most people instant silent results are achieved with... the LM4780 as a stereo amp.😎
You can figure out why.😉
Sheldon said:
You mean you are using a board made for the LM3875 with the LM3886?
Anyway, always glad to help.
Sheldon, with single PSU feeding two channels you need to get them as close as possible and make that thick wire joining the two channel's ground as short as possible. It's really a question of centimeters.
I don't have any noise with my LM3886 power amp (the mini-Krell😀 ), but I think that for most people instant silent results are achieved with... the LM4780 as a stereo amp.😎
You can figure out why.😉
Like this Uncle Carlos? 🙂
An externally hosted image should be here but it was not working when we last tested it.
Re: Re: Re: Re: Re: Re: Another T
Yes, I'm using the 3845 boards for the 3876 (long story involving smoke and local availability of the 3876 but not 3875). It did require a few modifications to account for the different pin arrangement. Also, this amp is inverted and the chip was designed for NI. Also, this amp has the T feedback network. However, the power pins are the same for both chips and the various traces are no longer than before. The chips are basically back to back with the heat sink in between. I may be able to shorten some of the connecting wires, but I'll wait until I get the stuff for the remote volume control, so I can do it all at one time.
Sheldon
carlosfm said:
Yes, I'm using the 3845 boards for the 3876 (long story involving smoke and local availability of the 3876 but not 3875). It did require a few modifications to account for the different pin arrangement. Also, this amp is inverted and the chip was designed for NI. Also, this amp has the T feedback network. However, the power pins are the same for both chips and the various traces are no longer than before. The chips are basically back to back with the heat sink in between. I may be able to shorten some of the connecting wires, but I'll wait until I get the stuff for the remote volume control, so I can do it all at one time.
Sheldon
- Status
- Not open for further replies.