DIGI 125 - right VAS selection

[Greg Erskine]

I apprepriate your replies cunningham, but I'm extremely busy at work and have 101 problems with my new house, so I hope to study your responses in detail on the weekend.

BTW: My version of winzip is having a problem with your zip files?? Something about version 2.1 ??

Thanks again.

[cunningham]

Originally posted by grege
I

BTW: My version of winzip is having a problem with your zip files?? Something about version 2.1 ??

Thanks again.

Download winzip version free trial from download.com

[Greg Erskine]

Download winzip version free trial from download.com

Work PC can't do.

I have the latest (and registered copy) at home.

[cunningham]

Originally posted by cunningham
Hi Grege,

In this circuit, D1 and D2 are 1A power diodes because they have a forward bias voltage greater than that of the base to emitter juction of outputs. ECG 185 type is the bias servo mounted with the outputs so that the temperature is the same. ECG185 type transistors have beta value close to MJ802 & MJ4502. Temperature coeffiecent is related to beta value. This allows you to adjust the bias on the outputs to just above cuttoff. measure the voltage accross the 0.1 Ohm emitter resistor, with no load. As soon as you see a few milivolts, your good.
The reason I used diodes between the bases of the outputs is explained in the Thevenin equivalent circuit for the AC model of a BJT. (see AC model.jpg on next post)
re is an internal component of a BJT. It is different depending on usually current rating, and gain. Basically, as the NPN transistor conducts more, the voltage on re and RE cause a voltage drop between the output and the NPN base. The emitter of the PNP output is at the potential of the load, or Vout, but the PNP base is pulled up by the two dioes D1 & D2 which drives the PNP output further into cut-off, which is good. As the signal crosses 0V, the voltage between the NPN base and emitter is back to 0.6V because the change in Vbe is porportional to the AC signal. Doing this will make your outputs happy.

The dependent current source (the collector) is dependent on base current and Beta.

By adding an emmitter resistor and base resistor, matching up transistors is not as important because re is not as big of a factor. The percent difference of re is larger than the percent difference of re + RE. True that adding an emitter resistor will reduce the gain, but not having to match componants is much cheaper.

You would drive this circuit with a VAS output of 0VDC and an impeadence of about 6-7000Ohms and a voltage of 23VRMS. By mounting ECG 54 & 55 on two separate 5W heatsinks, their temp. doesn't change very much so thier temperature coefficient is not a factor in the biasing, adding to stability. This is not a true darlington connection because the diodes bias the emitters of the drivers. Usually darlington pairs come in an inegrated package and are perfectly matched transistors inside and of course are always the same temp.

You can use any general purpose transistors for the current sources as long as they can handle 80V. only 100mW though so you might keep them small.

You can drive this circuit pretty hard and it is not likely that you will have any thermal runaway problems from it as long as your heat sink can sink 100W!!

Hope this helps a little.

[cunningham]

AND

[Greg Erskine]

cunningham,

I now have a printed copy so I can study them on the train.

Thanks