Re: ?
Hi, Jen and all,
It doesn't need to be. If you mount the 4 diodes on a single device, they sense more this transistor than the others. Re-read this part from prof. Marshall FAQ:
" The four bias diodes should see the average temperature of all four output transistors. This occurs only when all transistors and all bias diodes are on the same heat sink. "
The diodes aren't on the heat sink. You said it weighs 7kg, it'll take some time to all the output transistors share the same temperature, when needed.
You could measure the temperatures on each device, just to see if it really matters.
Hi, Jen and all,
JensRasmussen said:
It doesn't need to be. If you mount the 4 diodes on a single device, they sense more this transistor than the others. Re-read this part from prof. Marshall FAQ:
" The four bias diodes should see the average temperature of all four output transistors. This occurs only when all transistors and all bias diodes are on the same heat sink. "
The diodes aren't on the heat sink. You said it weighs 7kg, it'll take some time to all the output transistors share the same temperature, when needed.
You could measure the temperatures on each device, just to see if it really matters.
Placement of thermal diodes
I chose to put the diodes on the back of one transistor for several reasons.
I have two channels on one heatsink. By putting the diodes on the back of one transistor I ret a better reading for that amp.
Since the heatsink is heavy, I wanted to make sure that the output transisors did not overheat.
I have changed the Emitter resistors on the output stage to a bigger value than Leach uses in the original. This will help keeping the temperature for the different output transistors the same. (More feedback from a rise in current)
There might be (There problably is) better ways of thermal sensors, but I did not want to have the longer wires needed to measure the temperature of every transistor in the outputstage.
I'm thinking of using output transistors instead of diodes to measure the temperature of the heatsink (as a consequense of the descussion in this thread)
How could one measure the temperature of every outputtransistor? If you measure the case, you can only estimate the temperature of the semiconductor inside. And therefor the only use for the measured temperature is comparing to the heatsinkteperature, witch for for longer use is bound to be the same. I wanted to make sure that the outputdevices are not overheated before "Thermal Steady state"
\Jens
I chose to put the diodes on the back of one transistor for several reasons.
I have two channels on one heatsink. By putting the diodes on the back of one transistor I ret a better reading for that amp.
Since the heatsink is heavy, I wanted to make sure that the output transisors did not overheat.
I have changed the Emitter resistors on the output stage to a bigger value than Leach uses in the original. This will help keeping the temperature for the different output transistors the same. (More feedback from a rise in current)
There might be (There problably is) better ways of thermal sensors, but I did not want to have the longer wires needed to measure the temperature of every transistor in the outputstage.
I'm thinking of using output transistors instead of diodes to measure the temperature of the heatsink (as a consequense of the descussion in this thread)
How could one measure the temperature of every outputtransistor? If you measure the case, you can only estimate the temperature of the semiconductor inside. And therefor the only use for the measured temperature is comparing to the heatsinkteperature, witch for for longer use is bound to be the same. I wanted to make sure that the outputdevices are not overheated before "Thermal Steady state"
\Jens
Re: Placement of thermal diodes
Hi, Jens and all
Try a digital thermometer, or a temperature sensor and your multimeter. It's just for testing purposes, of course, you don't have to leave it in your amp forever, controling the output current...
(I'm sorry if I couldn't express myself properly before, english is not my native language)
A general purpose silicon diode can perform a good temperature sensor: you don't need to know the exact value, but to compare between several values.
Glue a diode on each tranny case and measure de DC voltage drop across the diodes, one at a time (you'll need a digital multimeter with a DIODE TEST range, usually indicated by a little diode symbol). They exhibit a negative temperature coefficient: the voltage drop decreases as temperature increases.
It can be useful to measure the heatsink too, and trace curves, just 4 fun.
And that's what you need, nothing more.
I had suggested you to measure each device, just to see if they could show unequal temperatures while the heatsink is warming up, or if (when) some unbalance occurs. It's only a guess.
Compare the temps of each tranny case at different working conditions. If they aren't the same (between them) all the time, you'll need to find a better way.
hope this helps,
Hi, Jens and all
JensRasmussen said:
Try a digital thermometer, or a temperature sensor and your multimeter. It's just for testing purposes, of course, you don't have to leave it in your amp forever, controling the output current...
(I'm sorry if I couldn't express myself properly before, english is not my native language)
A general purpose silicon diode can perform a good temperature sensor: you don't need to know the exact value, but to compare between several values.
Glue a diode on each tranny case and measure de DC voltage drop across the diodes, one at a time (you'll need a digital multimeter with a DIODE TEST range, usually indicated by a little diode symbol). They exhibit a negative temperature coefficient: the voltage drop decreases as temperature increases.It can be useful to measure the heatsink too, and trace curves, just 4 fun.
And that's what you need, nothing more.
I had suggested you to measure each device, just to see if they could show unequal temperatures while the heatsink is warming up, or if (when) some unbalance occurs. It's only a guess.
Compare the temps of each tranny case at different working conditions. If they aren't the same (between them) all the time, you'll need to find a better way.
hope this helps,
Hi Jens,
In the book by Self, he describes a situation very similar to yours. When he moved his temperature sensor from the heat sink to the transistor case, his output stage current became overcompensated. His fix for this was simple. He just increased the thermal resistance between the transistor case and the sensing element (a transistor in his design) with a shim. This made the temperature of the sensing device change less for a given junction temperature change and brought the amp back to optimum compensation. The designs by Self only sense the temperature of a single device, and he gets very good temperature compensation using 0.1 Ohm emitter resistors. He shows plots of bias current vs time after driving the amp hard, then removing the AC signal.
In the book by Self, he describes a situation very similar to yours. When he moved his temperature sensor from the heat sink to the transistor case, his output stage current became overcompensated. His fix for this was simple. He just increased the thermal resistance between the transistor case and the sensing element (a transistor in his design) with a shim. This made the temperature of the sensing device change less for a given junction temperature change and brought the amp back to optimum compensation. The designs by Self only sense the temperature of a single device, and he gets very good temperature compensation using 0.1 Ohm emitter resistors. He shows plots of bias current vs time after driving the amp hard, then removing the AC signal.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.