Would be better to match the Vbe as well within a gender. For even current sharing, matching across genders is not required, though it would be nice for better distortion and DC offset.
I bought power transistors in 25x tubes (On-semi) and found they were quite tightly close in Hfe and Vbe. Had no problems getting 4 groups of 5 out of the 25 with Hfe within 10% within a group, and the Vbe of all 25 in a tube within 2 or 3 mV (at about Ic=100mA). The Vbe was so tight that I had to measure it with an oscilloscope and pulse the Ic to catch the initial value before the self heating from the Ic would have drifted the Vbe.
One more thing - hFE matching does not make too much sense in this particular case.
If you want more even current distribution between the paralleled transistors - match transconductance.
Bias is voltage. For even Ic distribution, you need to group the transistors in a way that they will show close Ic values within the group of 5 transistors, having the same Vbe applied (just as you see it now).
P.S. Nattawa - typed simultaneously, basically the same thing
If you want more even current distribution between the paralleled transistors - match transconductance.
Bias is voltage. For even Ic distribution, you need to group the transistors in a way that they will show close Ic values within the group of 5 transistors, having the same Vbe applied (just as you see it now).
P.S. Nattawa - typed simultaneously, basically the same thing
Last edited:
Exactly, in an EF. Unless of course, in an unlikely case that a high value base stopper resistor is used that would translate Hfe variations into Vbe variation in a scale that can't be neglected.
Last edited:
Good enough in the circumstances we have.
Although, in my 2 x 5 pairs amplifier with IRFP240/9240 HexFETs, having roughly matched Vgs, the difference stays within 2mV throughout all pairs.
So, at some point over the time you may want to get some Vbe-matched transistors and exchange them for better current distribution, resulting in lower overall distortion.
Although, in my 2 x 5 pairs amplifier with IRFP240/9240 HexFETs, having roughly matched Vgs, the difference stays within 2mV throughout all pairs.
So, at some point over the time you may want to get some Vbe-matched transistors and exchange them for better current distribution, resulting in lower overall distortion.
Spooky grounds
I have traced out the grounds on the Jason Kuetemann Spooky PCB, Post 2069.
I have drawn the "differences" in red.
I have traced out the grounds on the Jason Kuetemann Spooky PCB, Post 2069.
I have drawn the "differences" in red.
in a push pull output stage the NPN to PNP does not matter much. the NFB adjusts the signal sent to the output stage to correct the error caused by the different upper and lower devices.
Matching of paralled devices is vital to avoid the excessive heating you have on the 18.6mV device compared to the average.
hFE matching alone is not sufficient.
Output current depends on the Vbe of the device at the operating current.
It is Vbe that needs to be matched AT the operating current you have chosen for the devices.
Last edited:
Curve tracer
I purchased this inexpensive curve tracer kit. Transistor Curve Tracer adapter Oscilloscopes Plus Power supply UnAssembled Kit | eBay
Surprisingly the cheap meters that allow hfe measurements are close.
I purchased this inexpensive curve tracer kit. Transistor Curve Tracer adapter Oscilloscopes Plus Power supply UnAssembled Kit | eBay
Surprisingly the cheap meters that allow hfe measurements are close.
Hi Thimios,
I would be interested in this as well as I cannot do the test myself.
Regards
Harry
In my first Slewmaster this things happens when I turn the power down: A very weak tone on a high frequency( I guess around 5kHz) start, but get weaker and weaker. On the power unit there are bleeder mounted, and I feel that the tone die out together with the emting of the power bank. Any idea?
Eivind S
Eivind S
I wonder how many people use the Spooky to drive subs and if they are used for general amplification on normal speakers. I now OS said they have more bass and are better for subs, but they should be good for driving normal speakers ?
ps Thimios, thanks for the experiment !
Hi Jan,
there are many that use the Spooky for their main amp as they like the sound.
OS does as well. I think for his sub he uses the Wolverine but I cannot find the post.
To do a fair comparison we need to sort out the "grounds" for the Spooky.
When the rail voltage goes way down while reservoir caps drain the storage, there could be a moment the front end stages drop out of their designed working conditions far enough, and from that moment on the feed back loop no longer have sufficient phase margin to stay stable. An oscillation would start. That is probably what you heard. This behavior is part of the reasons many speaker protection designs would cut out speakers early at power off, and delay speaker connection at power on.....So that you would not hear it.
I wonder how many people use the Spooky to drive subs and if they are used for general amplification on normal speakers. I now OS said they have more bass and are better for subs, but they should be good for driving normal speakers ?
ps Thimios, thanks for the experiment !
I modify Spooky amplifier for 90VDC PSU, the high frequency quality is good. I use slightly different compensation to achieve more slew rate.