Hi Clem,
Of course you can. That's why I said "We can't simulate that yet unless you want to take an RF viewpoint. That will keep you busy."
I just didn't put that in terms you'd recognise I guess.
So how are you? Got your's working yet?
Hi keantoken,
-Chris
Of course you can. That's why I said "We can't simulate that yet unless you want to take an RF viewpoint. That will keep you busy."
I just didn't put that in terms you'd recognise I guess.So how are you? Got your's working yet?
Hi keantoken,
Oh man! You would be a busy fella for sure. I'll bet this is done for high reliabilty boards, but not for hobby consumer grade stuff. Just by matching components you will go much further than any production piece. Just relax and get the simple stuff working. Accept the limitations of what you are doing.
-Chris
Well, I'm sure that LTSpice could integrate an RFI/EMI thing, but they might just opt not to because the sim was meant for PSU's. I'm sure that someone on DIYAudio could do something like that well, though...
It would be of great service and the creator would be honored throughout DIYAudio history... Maybe they'd even make a plaque for it!
Wouldn't it be great to have your own plaque hanging on the walls of this great DIYAudio website for all to see and admire? Nah... No one wants to, or do they? Any takers? I'd be a great opportunity...
It would be of great service and the creator would be honored throughout DIYAudio history... Maybe they'd even make a plaque for it!
Wouldn't it be great to have your own plaque hanging on the walls of this great DIYAudio website for all to see and admire? Nah... No one wants to, or do they? Any takers? I'd be a great opportunity...
anatech said:Hi Clem,
Of course you can. That's why I said "We can't simulate that yet unless you want to take an RF viewpoint. That will keep you busy."
So how are you? Got your's working yet?
Hi Chris,
Oh, ok. It's just that one wouldn't use SPICE to simulate RF in the first place, that's all.
I'm having problems sourcing the metal film 2W resistors - what's the group think on replacing them with standard carbon film 5% stuff?
Cheers
Clem
Keantoken, the bd139 in the to126 case was only chosen as it is a thermal tracking device and needs to be mounted on the heatsink.
About simulating board behaviour in sims, you might be able to add some properties like resistance and inductance of the routes, but it gets very difficult to also add the inductive and capacitive coupling between the routes. And it keeps questionable if it will ever really fit real world behaviour...
You need to adjust the pot so that the outputbias is the wanted ~60ma.
Kean, why not first understanding the fundamental electronics before trying the exotic stuff ?
About cancelling harmonics, that works, but is very difficult because the extra generated harmonics to subtract need to be a really good match, otherwise you only increase distortions.
Clem, carbon resistors should be no problem. It does not hurt if the 1.2ohm basestoppers are not 2W. Under normal operation they dissipate <100mw.
Mike
About simulating board behaviour in sims, you might be able to add some properties like resistance and inductance of the routes, but it gets very difficult to also add the inductive and capacitive coupling between the routes. And it keeps questionable if it will ever really fit real world behaviour...
You need to adjust the pot so that the outputbias is the wanted ~60ma.
Kean, why not first understanding the fundamental electronics before trying the exotic stuff ?
About cancelling harmonics, that works, but is very difficult because the extra generated harmonics to subtract need to be a really good match, otherwise you only increase distortions.
Clem, carbon resistors should be no problem. It does not hurt if the 1.2ohm basestoppers are not 2W. Under normal operation they dissipate <100mw.
Mike
Morning Clem!
For resistors, I used metal oxide in high power locations. That means those 1R2 resistors as well. I very rarely use carbon comps anymore but I'll have to get some for grid stoppers I guess.
-Chris
It would help if I knew what I was talking about, wouldn't it? I have never used a spice program, I guess once I have my testing and work 'puters squared away I should start modeling some simple circuits and get use to this. I'd much rather draw the schematic on paper and go. An old approach from the 60's and 70's I guess.Oh, ok. It's just that one wouldn't use SPICE to simulate RF in the first place, that's all.
For resistors, I used metal oxide in high power locations. That means those 1R2 resistors as well. I very rarely use carbon comps anymore but I'll have to get some for grid stoppers I guess.
-Chris
If you want to obtain a very good offset and a lowest distorsion spectrum you must match all the peers in two different conditions: at room temperature and heated by your hand (at appr 35-37grdC). In this way, you ashure that all peers devices have the same dhfe/dt behaviour.
All peers must be phisically coupled with thermal compound.
All peers must be phisically coupled with thermal compound.
Hi roender,
I think that's a bit beyond what needs to be done. If you look at my pictures you will see they are all in pairs, physically touching wth heatsink grease in between. I normally use heatshrink tubing over as well and will when I'm done experimenting.
If you buy transistors from the same batch and match the hFE while they are at the same temperature, they will track within what is required. Everyone can relax. Otherwise we'd all be having trouble with DC offset wandering all over the place.
-Chris
I think that's a bit beyond what needs to be done. If you look at my pictures you will see they are all in pairs, physically touching wth heatsink grease in between. I normally use heatshrink tubing over as well and will when I'm done experimenting.
If you buy transistors from the same batch and match the hFE while they are at the same temperature, they will track within what is required. Everyone can relax. Otherwise we'd all be having trouble with DC offset wandering all over the place.
-Chris
As a ex power semiconductor devices designer, in MESA diffusion departament at Baneasa SA, I can tell you that: dhfe/dt on the same silicon wafer can have 10-15% variation. One lot can have 10-100 different wafers.
I know that for diyers is not so important but if you search for "perfection" ...
I know that for diyers is not so important but if you search for "perfection" ...
Symasym does have one thermal problem pair: Q3/9
Q7 gets quite warm (~40°C), but is no issue as q8 defines the vbe needed. Thermal coupling between Q7/8 might even be a bad idea because of this.
Symasym's DC-offset does drift a bit, it measures <1mv when amp is cold, ~10 seconds later it reaches -3mv. (With no thermal coupling)
In theory you need also matching for following resistor pairs:
R5/6, R15/17, R14/29
Mike
Q7 gets quite warm (~40°C), but is no issue as q8 defines the vbe needed. Thermal coupling between Q7/8 might even be a bad idea because of this.
Symasym's DC-offset does drift a bit, it measures <1mv when amp is cold, ~10 seconds later it reaches -3mv. (With no thermal coupling)
In theory you need also matching for following resistor pairs:
R5/6, R15/17, R14/29
Mike
Hello Chris,
Make yourself the experiment that I propose to pinkmouse on 10 2n5551 or better on 10 MPSA18 (the effect is more visible at higher hfe) and note on a piece of paper the hfe values at 18-20grdC, room temperature, and 35-37grdC body temperature (keeping tight transistor capsule between your fingers).
Use folowing formula to calculate dhfe/dt for all your transistors from the "same lot".
[(hfe at Tbody) - (hfe at Troom)]/[Tbody - Troom]
Please let as know the results.
Mihai
Make yourself the experiment that I propose to pinkmouse on 10 2n5551 or better on 10 MPSA18 (the effect is more visible at higher hfe) and note on a piece of paper the hfe values at 18-20grdC, room temperature, and 35-37grdC body temperature (keeping tight transistor capsule between your fingers).
Use folowing formula to calculate dhfe/dt for all your transistors from the "same lot".
[(hfe at Tbody) - (hfe at Troom)]/[Tbody - Troom]
Please let as know the results.
Mihai
Hi Mihai,
Yes, the hFE will change. What is important is only how close they are to each other at different temperature points. How well they track.
So the experiment must be done with all transistors mounted in a heatsink along with a thermometer. Takes readings at several sections. Once stable, measure the hFE. Repeat for different temperatures. I'm going to bet that the matched pairs remain reasonably matched with temperature change.
This experiment is a pain, I've done it with output transistors. I've also done this with pre-matched pairs, but expect the same behaviour with pairs matched at one temperature.
-Chris
Yes, the hFE will change. What is important is only how close they are to each other at different temperature points. How well they track.
So the experiment must be done with all transistors mounted in a heatsink along with a thermometer. Takes readings at several sections. Once stable, measure the hFE. Repeat for different temperatures. I'm going to bet that the matched pairs remain reasonably matched with temperature change.
This experiment is a pain, I've done it with output transistors. I've also done this with pre-matched pairs, but expect the same behaviour with pairs matched at one temperature.
-Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.