Hi,
i have been testinga new power amp i got yesterday, the fostex 600.
It all works fine and i was rebiasing it when i came accross a possible problem, the bias seems to drift.
Its supposed to be between 5mV and 10mV accross 0.47ohms.
It starts at less than a millivolt then creeps up at a constant rate up until 8mV then i turned the amp off thinking somthings wrong.
It does this on both channels from power up over the course of a few minutes to reach 8mV.
I have posted this in the fostex 300 thread but i dont think many people read it.
Heres a link to the schematic post:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=79232&highlight=fostex+600
whats going on? is it normal?
Regards
Craig
i have been testinga new power amp i got yesterday, the fostex 600.
It all works fine and i was rebiasing it when i came accross a possible problem, the bias seems to drift.
Its supposed to be between 5mV and 10mV accross 0.47ohms.
It starts at less than a millivolt then creeps up at a constant rate up until 8mV then i turned the amp off thinking somthings wrong.
It does this on both channels from power up over the course of a few minutes to reach 8mV.
I have posted this in the fostex 300 thread but i dont think many people read it.
Heres a link to the schematic post:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=79232&highlight=fostex+600
whats going on? is it normal?
Regards
Craig
aha, good suggestion thanks! the bias has stopped drifting!
I kept it at less than 1mV per channel whilest it stabilized.
My only other concern with this amplifier is the heat generated by TR606 and TR605
(the ones in the diagram controlling current through the string of 5 diodes)
They get pretty hot, i probably had the amp on for 5 minutes for the bias stabilization and these transistors on both channels became very hot, i could hold my fingers on them but it was uncomfortable.
is this expected? only 10mA flowing through them at +/-85v rails (diagram says 75 but for some reason mine is 85, i guess the mains is dead on 240V here).
Thanks
Craig
I kept it at less than 1mV per channel whilest it stabilized.
My only other concern with this amplifier is the heat generated by TR606 and TR605
(the ones in the diagram controlling current through the string of 5 diodes)
They get pretty hot, i probably had the amp on for 5 minutes for the bias stabilization and these transistors on both channels became very hot, i could hold my fingers on them but it was uncomfortable.
is this expected? only 10mA flowing through them at +/-85v rails (diagram says 75 but for some reason mine is 85, i guess the mains is dead on 240V here).
Thanks
Craig
Thanks again,
They do have clip on heatsinks, i might just look into obtaining some larger ones for them to be on the safe side.
I figured out why my rails are high.
Its because the rails of 75V DC are obtained by running the amplifier off 220V AC, the amplifier is specified between 220 and 240v AC for operation.
I am on 240V from the socket but have been mostly testing at 220 using a variac so when i plug this thing straight in its rails are gonna be +/-85V which is gonna make things even hotter.
I think im going to run this amp from a variac permanently
edit:
Teh data sheet for these transistors specifies Max VCE of 150V so exceeding +/-75V would be putting the transistors under serious risk of exploding
Thanks
Craig
They do have clip on heatsinks, i might just look into obtaining some larger ones for them to be on the safe side.
I figured out why my rails are high.
Its because the rails of 75V DC are obtained by running the amplifier off 220V AC, the amplifier is specified between 220 and 240v AC for operation.
I am on 240V from the socket but have been mostly testing at 220 using a variac so when i plug this thing straight in its rails are gonna be +/-85V which is gonna make things even hotter.
I think im going to run this amp from a variac permanently
edit:
Teh data sheet for these transistors specifies Max VCE of 150V so exceeding +/-75V would be putting the transistors under serious risk of exploding
Thanks
Craig
Hi,
I think I understand how your circuit works.
You have a voltage (across r612//r614) feeding the base emitter of tr5 and this controls the current through tr6 and the variable resistor as well as the diode string.
There appears to be no constant current control on the voltage feeding the base nor temperature compensation on either the r612 voltage nor the diode+variable resistor.
Are the diodes in contact with some other semiconductor or heatsink?
The current through the VAS (tr6) and the current through R612//r614 both vary with rail voltage.
Increase your rails and the currents increase. So a 13% voltage increase will give rise to a 26% dissipation increase.
With this style of uncompensated VAS current any rail voltage variations will feed through into the output signal causing much increased distortion.
You have a three stage darlington output arrangement. With 3Vbe drops the stage bias is very susceptable to small changes in temperature of the devices. Combined with the variable and possibly uncompensated bias voltage coming from the diodes, then variations in Iq are VERY likely as the amp warms up.
The VAS and the Current source will both feel about 90% of rail to rail voltage when a high level signal at or near clipping is passed through the amp. So your 150V transistors will feel about 0.9*170V across them.
Have a look at the Tc derating for these transistors and at the SOAR at about 85Vce.
Apply both factors to your transistors and compare your permissible to the actual quiescent current.
Without checking the design, I suspect your transistors are in real stress when the rails are above +-75V and may be near their limit even at +-75V.
I think I understand how your circuit works.
You have a voltage (across r612//r614) feeding the base emitter of tr5 and this controls the current through tr6 and the variable resistor as well as the diode string.
There appears to be no constant current control on the voltage feeding the base nor temperature compensation on either the r612 voltage nor the diode+variable resistor.
Are the diodes in contact with some other semiconductor or heatsink?
The current through the VAS (tr6) and the current through R612//r614 both vary with rail voltage.
Increase your rails and the currents increase. So a 13% voltage increase will give rise to a 26% dissipation increase.
With this style of uncompensated VAS current any rail voltage variations will feed through into the output signal causing much increased distortion.
You have a three stage darlington output arrangement. With 3Vbe drops the stage bias is very susceptable to small changes in temperature of the devices. Combined with the variable and possibly uncompensated bias voltage coming from the diodes, then variations in Iq are VERY likely as the amp warms up.
The VAS and the Current source will both feel about 90% of rail to rail voltage when a high level signal at or near clipping is passed through the amp. So your 150V transistors will feel about 0.9*170V across them.
Have a look at the Tc derating for these transistors and at the SOAR at about 85Vce.
Apply both factors to your transistors and compare your permissible to the actual quiescent current.
Without checking the design, I suspect your transistors are in real stress when the rails are above +-75V and may be near their limit even at +-75V.
Thanks for the analysis AndrewT!
This citrcuit was designed around 1979 and uses old semiconducters which is probably why they run them close to their limits to achieve the output power (over 200w into 8 ohm). I am concerned about their voltage ratings too
, although this amplifier has been in use since about 1979 and has been unmodified as far as i can tell i.e. it has never blown up before.
The diodes in the bias network are mounted on a transistor heatsink i think, the diagrm shows them in a dotted box.
As the THD of the amplifier at its rated output is no more than 0.05% so the bias issues seem to have been accounted for.
No the amp doesnt have a 220/240v switch, it specifies it can be run at between 220 and 240.
Thanks a lot
Craig
This citrcuit was designed around 1979 and uses old semiconducters which is probably why they run them close to their limits to achieve the output power (over 200w into 8 ohm). I am concerned about their voltage ratings too
, although this amplifier has been in use since about 1979 and has been unmodified as far as i can tell i.e. it has never blown up before.The diodes in the bias network are mounted on a transistor heatsink i think, the diagrm shows them in a dotted box.
As the THD of the amplifier at its rated output is no more than 0.05% so the bias issues seem to have been accounted for.
No the amp doesnt have a 220/240v switch, it specifies it can be run at between 220 and 240.
Thanks a lot
Craig
Hi,
200W into 8ohm is a lot less stressful on output and driver stage than 350W into 4ohms and much less stressfull than 450W into an even lower load.
BTW 200 into 8ohms only requires a Vpk into load of 56Vpk. Most amps of about this power would run on +-65V to +-70V.
Have you tried measuring the supply rails when the amp is delivering significant power into a high resistance (8ohm)? I wonder what the 71V means beside the smoothing caps.
An amp that has survived unrepaired for over 25 years indicates a well balanced choice of components.
200W into 8ohm is a lot less stressful on output and driver stage than 350W into 4ohms and much less stressfull than 450W into an even lower load.
BTW 200 into 8ohms only requires a Vpk into load of 56Vpk. Most amps of about this power would run on +-65V to +-70V.
Have you tried measuring the supply rails when the amp is delivering significant power into a high resistance (8ohm)? I wonder what the 71V means beside the smoothing caps.
An amp that has survived unrepaired for over 25 years indicates a well balanced choice of components.
Amps that drift until thermal equilibrium is achieved, can be reliable. I have built amps like this and would periodically measure the bias current when warm. Even with moderate variations in ambient temperature, the bias current was predictable.richie00boy said:Yes again the proof of the pudding is in the eating. Now you have set bias I would just forget about it all
The circuit simplicity was not necessarily a bad thing IMO. If anything I think you should simply ensure a consistent airflow through the amp.
Thanks for replies everyone.
AndrewT,
I havent measured supplies under load yet, i wouldnt expect them to dip much, the transformer is massive (1.3KVA) plus 36000uf of capacitance.
I am waiting for another DIY member (thejohn) to verify my finding with his lab 600 amplifier which he has run for some time.
I would agree that if the amp hasnt been repaired ever in 25 years its got to be doing somthing right!
Indm,
I have set the bias to a steady 6.5mA after 20 mins of stabilising time, this way i think it should hit around 10mA when the case heats up. Although i will have to check this.
A fan would certainly help keep the transistors inside cooler, but i wonder if it would keep the bias from stabilising? When i biased this thing my window was open and the occasional draft would disturb the bias.
Regards
Craig
AndrewT,
I havent measured supplies under load yet, i wouldnt expect them to dip much, the transformer is massive (1.3KVA) plus 36000uf of capacitance.
I am waiting for another DIY member (thejohn) to verify my finding with his lab 600 amplifier which he has run for some time.
I would agree that if the amp hasnt been repaired ever in 25 years its got to be doing somthing right!
Indm,
I have set the bias to a steady 6.5mA after 20 mins of stabilising time, this way i think it should hit around 10mA when the case heats up. Although i will have to check this.
A fan would certainly help keep the transistors inside cooler, but i wonder if it would keep the bias from stabilising? When i biased this thing my window was open and the occasional draft would disturb the bias.
Regards
Craig
Hi,
you have either got to prove that the devices will continue to be reliable at +-85V
or
reduce the rail voltage to the manufacturer's specified +-75V.
I would not support the thought of leave well alone, partly for the reasons given earlier but mostly to find out why the voltage is apparently 13% above specification and it's effect on dissipation abilities, warm up time, ultimate device quiescent temperatures, device peak temperatures, and variation of all of these with changes in ambient temperature (room and case).
BTW. 36mF is very low for the current ability of this amp.
Using information from other designers and findings of my own I always recommend a minimum of 2mF per Apk of output current.
On that basis your 18mF can support a maximum of 4.5Apk per channel. This equates to a minimum load impedance of about 13r. If I were designing this amp for 4ohm duty I would be using +-40mF per channel (a total of 160mF not 36mF) and I would probably try increasing to +-60mF to listen for any improvements before releasing build information.
you have either got to prove that the devices will continue to be reliable at +-85V
or
reduce the rail voltage to the manufacturer's specified +-75V.
I would not support the thought of leave well alone, partly for the reasons given earlier but mostly to find out why the voltage is apparently 13% above specification and it's effect on dissipation abilities, warm up time, ultimate device quiescent temperatures, device peak temperatures, and variation of all of these with changes in ambient temperature (room and case).
BTW. 36mF is very low for the current ability of this amp.
Using information from other designers and findings of my own I always recommend a minimum of 2mF per Apk of output current.
On that basis your 18mF can support a maximum of 4.5Apk per channel. This equates to a minimum load impedance of about 13r. If I were designing this amp for 4ohm duty I would be using +-40mF per channel (a total of 160mF not 36mF) and I would probably try increasing to +-60mF to listen for any improvements before releasing build information.
If I remember correctly, a fan is not a complete substitute for a heatsink, but is still a good thing. A fan does reduce sensitivity to ambient conditions, stabilising the airflow by outweighing natural air movements.
I used to use a couple of computer fans running on maybe 6 or 7 volts so they were more quiet.
I used to use a couple of computer fans running on maybe 6 or 7 volts so they were more quiet.
Craig,
So I took the cover off my 600 today to see if I could verify your findings so far...
With no load connected and gains set all the way down as per the service manual, voltmeter on R634 (one of the 0.47ohm output device emitter resistors), switch on and....
Pretty much 0mv! It then climbed slowly to 8mv after about 10min and settled there. TR605 and 606 heatsinks barely warm.
Powered down, connected load, on again, play some music, amp nice and warm, feel TR506 and 606 again... warm not hot (maybe 40something deg?)
Disconnect load, gains down, check bias, amp still nice and warm... still steady at 8mv.
My mains voltage today is sitting at 224v. We are meant to have 230 but in reality it varies between 220 and 240!
My rail voltages, measured on the main psu caps are +-74v.
Temperature today here is about 24deg. In summer we have regular 30deg+ ambient, I haven't checked bias under these conditions...
Hope this helps...
John
So I took the cover off my 600 today to see if I could verify your findings so far...
With no load connected and gains set all the way down as per the service manual, voltmeter on R634 (one of the 0.47ohm output device emitter resistors), switch on and....
Pretty much 0mv! It then climbed slowly to 8mv after about 10min and settled there. TR605 and 606 heatsinks barely warm.
Powered down, connected load, on again, play some music, amp nice and warm, feel TR506 and 606 again... warm not hot (maybe 40something deg?)
Disconnect load, gains down, check bias, amp still nice and warm... still steady at 8mv.
My mains voltage today is sitting at 224v. We are meant to have 230 but in reality it varies between 220 and 240!
My rail voltages, measured on the main psu caps are +-74v.
Temperature today here is about 24deg. In summer we have regular 30deg+ ambient, I haven't checked bias under these conditions...
Hope this helps...
John
AndrewT,
I have decided to run my amp off off 220v from a variac permanently to avoid trouble, i like to feel safe when listening to music. It kind of ruins the experience if all you can think of is transistors getting hotter and hotter then boom!
Thejohn, thanks for checking your amp for me.
The fact thast your heatsinks are only warm concerns me, i have checked the voltage measurements around these transistors and they are exactly as the schematic specifies... as are all other voltages marked for measurement?
I dont understand how they could heat up so much if yours are not
Do dying transistors get hotter? i have no idea, but the PNp and the NPN get equally hot on each channel.
Im going to have a serious probe after the football and see if i can find a problem.
Thanks
Craig
I have decided to run my amp off off 220v from a variac permanently to avoid trouble, i like to feel safe when listening to music. It kind of ruins the experience if all you can think of is transistors getting hotter and hotter then boom!
Thejohn, thanks for checking your amp for me.
The fact thast your heatsinks are only warm concerns me, i have checked the voltage measurements around these transistors and they are exactly as the schematic specifies... as are all other voltages marked for measurement?
I dont understand how they could heat up so much if yours are not
Do dying transistors get hotter? i have no idea, but the PNp and the NPN get equally hot on each channel.
Im going to have a serious probe after the football and see if i can find a problem.
Thanks
Craig
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.