Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Solid State Talk all about solid state amplification.

How do you match tempco between spreader and power transistor?
How do you match tempco between spreader and power transistor?
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Closed Thread
 
Thread Tools Search this Thread
Old 4th August 2015, 01:10 AM   #1
Alan0354 is offline Alan0354  United States
Banned
 
Join Date: Apr 2014
Location: US
Default How do you match tempco between spreader and power transistor?

I am working on a 3EF diamond OPS and the bias spreader is only 2Vbe+52mV. This rules out a lot of designs covered in the books and posts.

I am using CFP bias spreader using a PNP Q18 to help the NPN Q10 bias spreader. I originally use KSC3503 as Q10 bolting on top of the big power transistor, using BC560 as PNP Q18. Since it's 3EF Diamond, the spreader voltage is down to 1.1V only. The divider ratio is way less than 2 and I don't get -4mV/deg C ( just assume -2mV/deg C per Vbe for now). I went into thermal run away. The Vbe KSC was about 0.66V, it's too high, so I choose using MJW big transistor as the spreader to get the divider ratio better because the Vbe is only 0.51V. That works perfect for the CFP with R37 equal 270ohm.

I want to understand this a little more. I followed Bonsai's paper that adjusting R37 will change the tempco. So I change R37 to a adjustable pot to adjust from 0 to 1K. I adjust from 0 to some value. This is my observation:

1) The value of R37 and the TPOT1 are very interactive. You cannot adjust one without the need to adjust the other to keep the bias current constant.

2) When R37 is 0, this remove the PNP Q18 from the picture and the spreader becomes a simple spreader with a single NPN transistor that is common in a lot of circuit. With this I slowly get thermal run away even with the big MJW as spreader transistor. This is understandable because the total voltage is 1.1V, way less than 2Vbe and therefore the tempco is less than -4mV/deg C.

3) When I increase R37 above 270ohm as shown in the schematic, I have negative current tempco.......meaning the current decrease as the transistors heat up.

4) decreasing R37 reduce the negative tempco of the spreader.

5) With the MJW that Vbe=0.51V, the collector base voltage is running between -0.16 to -0.25V. The transistor is still running in linear mode. If using a transistor that has over 0.6V, the transistor will be much closer to saturation. So finding a transistor with low Vbe is important.

So this tells me I need to pick the right transistor, the right value for R37. This will change with how good the thermal coupling from the power transistor and the spreader, it also depends on the consistency of the Vbe of the spreader transistor even you use the same transistor, they still can vary from device to device.


My question to you that are more experience, how do you design the spreader? How do you adjust the tempco. Any way to design this without resort to try and true to find the right combination like what I did?

Is this short coming only for 3EF Diamond? Because if you have a 4Vbe or 6Vbe spreader, there are so many options to compensate.

Thanks
Attached Images
File Type: jpg CFP bias spreader.jpg (22.7 KB, 388 views)

Last edited by Alan0354; 4th August 2015 at 02:03 AM.
 
Old 4th August 2015, 02:13 AM   #2
nattawa is offline nattawa  Canada
diyAudio Member
 
Join Date: Apr 2010
Location: Mississauga, Ontario
for 2x Vbe spreading make R10=(R18+POT), trim R37 to arrive at desired OPS idle current. R37 may be much greater than 270R depending on what is being used as Q9. By doing so you are hunting an Ic of Q9 where its Vbe matches that of the OPS trannies.

I have used this spreader circuit with my ThermalTrak OPS and it has been rock solid with heat sinks tracking the ambient temp at almost a constant rise winter or summer. I use MJE340 for Q9 and it`s responsible for tracking pre-drivers and drivers in a 3EF.
 
Old 4th August 2015, 02:53 AM   #3
Alan0354 is offline Alan0354  United States
Banned
 
Join Date: Apr 2014
Location: US
Quote:
Originally Posted by nattawa View Post
for 2x Vbe spreading make R10=(R18+POT), trim R37 to arrive at desired OPS idle current. R37 may be much greater than 270R depending on what is being used as Q9. By doing so you are hunting an Ic of Q9 where its Vbe matches that of the OPS trannies.

I have used this spreader circuit with my ThermalTrak OPS and it has been rock solid with heat sinks tracking the ambient temp at almost a constant rise winter or summer. I use MJE340 for Q9 and it`s responsible for tracking pre-drivers and drivers in a 3EF.

Hi Nattawa

Thanks, I'll try making R10=R18+POT and adjust the R37 and see.

Is making R10=R18+POT guaranty the tempco of 2Vbe, then use R37 to reduce down to the right voltage?

Do you know the Vbe of MJE340? I was looking at this one, but data sheet usually do not give the Vbe for a set current for medium power transistors, only the small signal transistors. KSC3503 has high Vbe.


Is this supposed to be common knowledge that everybody here supposed to know, I really have not seen post talking about this from all the posts I read here the last few months. As you know, I am new,iIt did not even come up on my radar until I saw the problem. It does not help that I need a 2Vbe spreader that a lot of the clever design will not work.

EDIT:

I just adjust to R10=R18+POT and adjust R37 to get 26mV across the Re. I'll post back the result.

Last edited by Alan0354; 4th August 2015 at 03:06 AM.
 
Old 4th August 2015, 03:27 AM   #4
Alan0354 is offline Alan0354  United States
Banned
 
Join Date: Apr 2014
Location: US
The current edge up very slowly. I stopped at voltage across Re rise from 27mV to 32mV. But is rise up slower and slower. It's very close.

I am increasing the value of the POT from 250ohm to 230ohm. I am cooling it down and start again.

Last edited by Alan0354; 4th August 2015 at 03:54 AM.
 
Old 4th August 2015, 03:29 AM   #5
nattawa is offline nattawa  Canada
diyAudio Member
 
Join Date: Apr 2010
Location: Mississauga, Ontario
Quote:
Originally Posted by Alan0354 View Post

Is making R10=R18+POT guaranty the tempco of 2Vbe, then use R37 to reduce down to the right voltage?

Do you know the Vbe of MJE340? I was looking at this one, but data sheet usually do not give the Vbe for a set current for medium power transistors, only the small signal transistors. KSC3503 has high Vbe.


Is this supposed to be common knowledge that everybody here supposed to know, I really have not seen post talking about this
Yes, tempco of 2x Vbe would be a starting point. Q18 would introduce a slight extra negative tempco and that could be welcome. The thermal thing does not usually become common knowledge before finishing one's first build.

Your 2x Vbe can be challenging as you have very little voltage room to play, Q9 is already in quasi saturation and its collector potential may be lower than what is on its base. And I don't know how this may affect the spreader performance.

I used to propose a 2x Vbe spreader using an op amp for diamond driver 3EF OPS in perhaps way back in Bob's book thread. I actually put that circuit to test and it worked. But I later changed mind deciding not to use a diamond driver.

I don't remember MJE340 Vbe particularly but it should be somewhere close to what ksc3503 has.
 
Old 4th August 2015, 04:04 AM   #6
Alan0354 is offline Alan0354  United States
Banned
 
Join Date: Apr 2014
Location: US
Quote:
Originally Posted by nattawa View Post
Yes, tempco of 2x Vbe would be a starting point. Q18 would introduce a slight extra negative tempco and that could be welcome. The thermal thing does not usually become common knowledge before finishing one's first build.

Your 2x Vbe can be challenging as you have very little voltage room to play, Q9 is already in quasi saturation and its collector potential may be lower than what is on its base. And I don't know how this may affect the spreader performance.

I used to propose a 2x Vbe spreader using an op amp for diamond driver 3EF OPS in perhaps way back in Bob's book thread. I actually put that circuit to test and it worked. But I later changed mind deciding not to use a diamond driver.

I don't remember MJE340 Vbe particularly but it should be somewhere close to what ksc3503 has.
The base of Q9 is about 200mV lower than the base already with the MJW. But I think it's still running in linear mode, not quite saturate yet. That's the problem with the 2Vbe, I don't think I can use a transistor with Vbe of 0.6V and over. But only very huge transistor like MJW will have low Vbe like this because they meant for high current, so the Vbe will reduce logarithmically as you use low current. That's the reason I picked the big MJW on the first place.

I am going to change to a lower value pots for R37 to 500ohm. And I ordered 200ohm trim pot so I can change the TPOT1 from 500 to 200ohm.

I don't know whether I can find an opamp that work with supply down to 1V, it's hard to use an opamp as bias spreader. I really don't want to have a floating supply to power the opamp.

Nobody warn about the 3EF diamond about the difficulty of spreader. But it is very stable, the tempco of the pre-driver and the driver really cancelled out very good.

Thanks for your help. I'll report back on further finding.

Last edited by Alan0354; 4th August 2015 at 04:21 AM.
 
Old 4th August 2015, 09:24 AM   #7
Alan0354 is offline Alan0354  United States
Banned
 
Join Date: Apr 2014
Location: US
I have been massaging the values, the best result is still very close to my original value. R37 is about 220ohm and R18+TPOT1=855ohm.

So the resistor divider for the spreader is 1+1000/855=2.17. Therefore the best result is with tempco of 2.17Vbe.

I next step is to try to replace the MJW to a smaller transistors. I need to find a T0-126 NPN that has low Vbe on. I also need to find a PNP for the CFP spreader.

Last edited by Alan0354; 4th August 2015 at 09:30 AM.
 
Old 4th August 2015, 10:29 AM   #8
AndrewT is offline AndrewT  Scotland
R.I.P.
 
Join Date: Jul 2004
Location: Scottish Borders
Quote:
Originally Posted by Alan0354 View Post
...........

Is making R10=R18+POT guaranty the tempco of 2Vbe, ..............
almost exact.
The two resistances are in 1:1 ratio.
Therefore the two voltages are in 1:1 ratio and one of these voltage is Vbeq9

But some current in the resistor string is stolen to feed the base of Q9.
This introduces an error in the 1:1 ratio.
The lower the hFE of Q9 the bigger the error. Conversely the higher the hFE the smaller the error from the exact 1:1 voltage ratio.
Quote:
resistor divider for the spreader is 1+1000/855=2.17. Therefore the best result is with tempco of 2.17Vbe.
This ratio needs to be corrected for the base current drawn from the resistor string.


I have used a common guideline from decades of tube/valve advice, when I use the single transistor Vbe multiplier.
The base current < 1/10th of the resistor current. Higher hFE gets base current << 1/10th resistor current.

The current passing through the Vbe multiplier affects the base current drawn from the resistor string. If the output stage draws a varying current then the multiplier current changes and that forces the base current to change by a small amount. This affects the set point of the Vbe multiplier. Remember the Vbe multiplier is a Shunt Regulator.

Go back to D.Self. His extra resistor was a good way of reducing changes in the set point when a variable current was drawn by the output stage.
There is a paper by another author that explains this in more detail. I don't know which came first.

Low Vbe of any transistor is determined by how little Ie is drawn. Low current density equals low Vbe. Almost every datasheet confirms this.
A big power transistor will be lower Vbe than a small signal transistor at the same Ie.
A big power transistor will be lower hFE than a small signal transistor at the same Ie.
These two could be used to select a more optimal Q9. I don't have suggestions.
__________________
regards Andrew T.

Last edited by AndrewT; 4th August 2015 at 10:39 AM.
 
Old 4th August 2015, 10:35 AM   #9
franzm is offline franzm
diyAudio Member
 
Join Date: Dec 2010
What about using thermistors in the divider?
 
Old 4th August 2015, 03:12 PM   #10
Pingrs is offline Pingrs  United Kingdom
diyAudio Member
 
Join Date: Jul 2005
Location: NW England
[QUOTE=AndrewT;4410552]
..There is a paper by another author that explains this in more detail.
/QUOTE]

This, perhaps?

Brian.
Attached Files
File Type: pdf hagerman_vbe.pdf (212.3 KB, 116 views)
 

Closed Thread


How do you match tempco between spreader and power transistor?Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
bias spreader bypass cap Dan Moos Solid State 6 22nd April 2013 04:52 AM
Help understanding bias spreader Dan Moos Solid State 7 5th March 2013 02:52 PM
Want to Buy: J50 Transistor (2pcs match) mrquery Swap Meet 4 28th October 2007 03:53 PM
decreasing tempco in this Vbe multipl. djdamix Solid State 5 26th February 2005 08:57 PM
Magnatec FET's and their negative tempco Petter Parts 3 1st December 2002 03:43 PM


New To Site? Need Help?

All times are GMT. The time now is 10:17 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Copyright ©1999-2019 diyAudio
Wiki