What would be an easy way to adjust DC offset in this circuit?
I know the typical way with most AB amps is a 100k pot at base of inverter, series 300k resistor at base and + & - 15v at each side of pot. Heres the schematic
http://www.munsonandbryan.com/400a.tif[/IMG]
Would this dc offset schematic work in circuit?
http://www.munsonandbryan.com/dc offset.jpg[/IMG]
I know the typical way with most AB amps is a 100k pot at base of inverter, series 300k resistor at base and + & - 15v at each side of pot. Heres the schematic
http://www.munsonandbryan.com/400a.tif[/IMG]
Would this dc offset schematic work in circuit?
http://www.munsonandbryan.com/dc offset.jpg[/IMG]
Hi John,
Okay, don't modify the circuit. Match the 6571 pair. The Hfe circuit on small digital meters will work fine for this. This will drop your DC offset to a low value and improve the sound. While you're in there, replace the 470uF 16V cap unless it's new. You can calculate the value of DC offset closely by dividing the tail current by two. Divide the figure by the Hfe of a 6571 transistor. Multiply this by 9400 plus the output DC resistance of the preamp. This unit does not have a DC reference to ground for the input. Not nice! You could hang a resistor from somewhere on the input to true ground (not the RCA jack ground).
If you were to unplug the source from the amp. the DC output voltage will rise from X mV to some undesireable level.
-Chris
Okay, don't modify the circuit. Match the 6571 pair. The Hfe circuit on small digital meters will work fine for this. This will drop your DC offset to a low value and improve the sound. While you're in there, replace the 470uF 16V cap unless it's new. You can calculate the value of DC offset closely by dividing the tail current by two. Divide the figure by the Hfe of a 6571 transistor. Multiply this by 9400 plus the output DC resistance of the preamp. This unit does not have a DC reference to ground for the input. Not nice! You could hang a resistor from somewhere on the input to true ground (not the RCA jack ground).
If you were to unplug the source from the amp. the DC output voltage will rise from X mV to some undesireable level.
-Chris
DJK says to replace 300ohm with 10turn 500ohm.
I experimented with paralleling pot, it took about 280ohm to increase offset by 10mv. If I do the opposite by increasing the resistance, it seems that it would take about 400-500ohms to get 0 offset. Would this not decrease the current to diff pair and change the general performance of the circuit.
I experimented with paralleling pot, it took about 280ohm to increase offset by 10mv. If I do the opposite by increasing the resistance, it seems that it would take about 400-500ohms to get 0 offset. Would this not decrease the current to diff pair and change the general performance of the circuit.
The offset is due to the difference in base currents in the 6571 pair. By increasing the 300 ohm resistance, you lower the tail current, and thus lower the base current for both transistors, and offset goes down. This isn't the way to solve the problem, as output drive for this stage decreases with decreased tail current.
Put in a pair of high-gain transistors, matched as closely as you can get them. High-gain NPN's are fairly easy to locate...perhaps a 2SC1775 or a Zetex ZTX694B (I've used the Zetex in this capacity with very good results).
With a tail current of about 2 1/2mA, and no provisions for adjustment, you will be lucky to get offset down below 20mV no matter how close you match transistors. But it is certainly about your only option besides radically changing circuit parameters (not recommended) or adding a servo (not a bad idea, but not as easy as maching transistors).
Put in a pair of high-gain transistors, matched as closely as you can get them. High-gain NPN's are fairly easy to locate...perhaps a 2SC1775 or a Zetex ZTX694B (I've used the Zetex in this capacity with very good results).
With a tail current of about 2 1/2mA, and no provisions for adjustment, you will be lucky to get offset down below 20mV no matter how close you match transistors. But it is certainly about your only option besides radically changing circuit parameters (not recommended) or adding a servo (not a bad idea, but not as easy as maching transistors).
As long as the DC offset is below 50mV, who cares? You won't damage the speakers until the offset gets much higher.
Echowars, with matched transistors, I routinely get the DC offset within a few mV of my calculated values for a given circuit design. It really isn't that difficult to do.
John, what is happening that causes you to want to reduce the DC offset? If you get a turn on/off thump, then there are transients that are causing this. You will not get a sound from a steady DC offset across your speakers at these magnitudes. If you plug and unplug your speakers withthe amp running (no sound), you may hear a faint click. That's all this does. Any change in the input impedance of this amp will cause a bigger noise than this offset. Did you calculate the expected DC offset of this amplifier?
-Chris
Echowars, with matched transistors, I routinely get the DC offset within a few mV of my calculated values for a given circuit design. It really isn't that difficult to do.
John, what is happening that causes you to want to reduce the DC offset? If you get a turn on/off thump, then there are transients that are causing this. You will not get a sound from a steady DC offset across your speakers at these magnitudes. If you plug and unplug your speakers withthe amp running (no sound), you may hear a faint click. That's all this does. Any change in the input impedance of this amp will cause a bigger noise than this offset. Did you calculate the expected DC offset of this amplifier?
-Chris
But you're not receiving the full benefit of the distortion cancelation abilities of the diff pair. We're tweakers, remember?anatech said:
0.00V offset is da goal.
Your meter is better than my meter. I bow to you Sir.
(thermal parameters also make this difficult, but U know that)Getting offset down as low as possible just happens to be a fetish with some of us. Don't necessarily have to have a reason, other than to know that 1mV is better than 100mV.
Hi Echowars,
Actually, a single diff pair will have a design positive or negative offset (NPN pair has negative offset). So the perfect balance is obtained at this theoretical value. Use the method I gave earlier to figure this value out for your particular design. Once the diff pair is balanced, you are achieving the full benefits of your diff pair.
Now, I like to see 0mV offset as much as the next person (don't ask me why, I'm a perfectionist I guess). This can be accomplished by injecting a correction current to offset the normal bias current of the diff pair. Just pick a side.
Thermal drift is minimized by thermally coupling the diff pairs and matching the other transistors in the amp. A well designed amp will then have minimal thermal drift with bias and offset values. I like to see the amp drift into balance rather than out.
I don't know that my meter is any different than yours. I buy 20 - 30 pcs of signal transistors at a time and match them at one sitting.
-Chris
Actually, a single diff pair will have a design positive or negative offset (NPN pair has negative offset). So the perfect balance is obtained at this theoretical value. Use the method I gave earlier to figure this value out for your particular design. Once the diff pair is balanced, you are achieving the full benefits of your diff pair.
Now, I like to see 0mV offset as much as the next person (don't ask me why, I'm a perfectionist I guess). This can be accomplished by injecting a correction current to offset the normal bias current of the diff pair. Just pick a side.
Thermal drift is minimized by thermally coupling the diff pairs and matching the other transistors in the amp. A well designed amp will then have minimal thermal drift with bias and offset values. I like to see the amp drift into balance rather than out.
I don't know that my meter is any different than yours. I buy 20 - 30 pcs of signal transistors at a time and match them at one sitting.
-Chris
Well, I toasted one of my speakers, the side that had 75mv offset.
Also killed a couple outputs. Will be changing all my outputs next week with MJ15024/25. Hopefully a smooth transistion as the SOA is higher than the 2N5876/78 but HFE is not the same.
I have matched the beta on new transistors which should help some. Changing all the lytics
Looks like I have some tweaking to do.
Also killed a couple outputs. Will be changing all my outputs next week with MJ15024/25. Hopefully a smooth transistion as the SOA is higher than the 2N5876/78 but HFE is not the same.
I have matched the beta on new transistors which should help some. Changing all the lytics
Looks like I have some tweaking to do.
Hi John,
75mV will not cause any speaker problems. There is something else going on here. The speaker died when the output stage failed, you probably had a large DC offset.
Your pick of output devices is fine. Hfe differences will not matter as there is a large range for each device.
When you repair this channel, replace the drivers and bias transistor as a minimum in addition to the outputs.
DJK, as I look back, I see I missed the 15K resistor. Thanks for pointing that out. THere's always room for a second set of eyes!
-Chris
75mV will not cause any speaker problems. There is something else going on here. The speaker died when the output stage failed, you probably had a large DC offset.
Your pick of output devices is fine. Hfe differences will not matter as there is a large range for each device.
When you repair this channel, replace the drivers and bias transistor as a minimum in addition to the outputs.
DJK, as I look back, I see I missed the 15K resistor. Thanks for pointing that out. THere's always room for a second set of eyes!
-Chris
I followed the advice from several people and ordered MJ15024/25 from ONSEMI via the sample program. I received 40 transistors from the same batch and by my suprise the beta was very close but I matched them up anyway before installing. I also replaced all the emitter resistors which were also handmatched. When the one channel blew it also took out a 1n4004 and 2-MPSL01 which I put a MPSA06 temporarly in its place at thermal circuit and a NTE287 in cascode. The MPSA06 was put in in the thermal circuit which appears to mute the input when thermal breaker opens. I am waiting on some black gates and other misc. parts but couldnt resist and had to test the transistors anyway. I ramped the amp up with my variac and the bias was rock solid. The output was about 118watts and the dc offset was about 35mv into a 8ohm 250w load resistor, no oscillation. My 400A should sing with the new caps, hookup wire, bridge rect. and resistors. I dont see the point in replacing the 2-680pf silver micas, could you explain or are you referring to the 390pf in the Stasis Circuit? I am having trouble sourcing computer grade caps that will fit this chassis. I need a 2-1/2" dia 100v cap and no taller than 4-1/2", the current caps are 15,000mfd. I tested them with a small load and they seem to hold a good charge but they may not be to spec. My ESR meter pegs out on these computer grade caps. I may add some bypass caps and call it a day.
I am also updating a SL-10 and will let you know how that project goes. All the caps in the SL-10 are labeled "Threshold" and do not show polarity or values. The polarity is not a problem but I will probably max out the electrolytic caps for the space and mounting.
Thanks
John
I am also updating a SL-10 and will let you know how that project goes. All the caps in the SL-10 are labeled "Threshold" and do not show polarity or values. The polarity is not a problem but I will probably max out the electrolytic caps for the space and mounting.
Thanks
John
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.