When replacing mosfet finals, should they be tested and matched, or can you just slap a couple in. Just curious.
-
You can now help support WorldwideDX when you shop on Amazon at no additional cost to you! Simply follow this Shop on Amazon link first and a portion of any purchase is sent to WorldwideDX to help with site costs.
Mosfet matching necessary?
- Thread starter AxlMyk
- Start date
I wish I could say yes to that - but if you have any capacitance checker you might want to look at Gate capacitance, compare that leg to the Source and Drain legs - then if the two parts (or how many you need) you are checking are within 5% tolerance of each other - I'd call it a pairing.
That is the only safer method I know of unless you trust the checkers for Siemens (mΩ) ratings - for that is also not considering the legs own inductance and ohmic alloy result in the epoxy package.
The Gate capacitance in RF strips in the more sensitive of the values to dropping in a replacement.
That is the only safer method I know of unless you trust the checkers for Siemens (mΩ) ratings - for that is also not considering the legs own inductance and ohmic alloy result in the epoxy package.
The Gate capacitance in RF strips in the more sensitive of the values to dropping in a replacement.
The reason I ask is one day my Q5n started losing power. I heard a snap and the 10a fuse blew. I ordered a pack of 13n10 mosfets, picked out 3, and replaced the driver and both finals. That fixed it. I put the pulled mosfets aside. A real cheap fix.
I bought one of those $22 testers on Amazon and used it to find the bad one. The driver had failed.
I used the tester to check all my 13n10, 13n10L, and IRF520 mosfets. There were some definite differences in them. Not large, but noticeable.
C readings of the 13n10s were 712 to 718 pF, with Vt from 3.9 to 4.1v.
The ones that were removed read 712 and the lowest Vt at 3.7.
I bought one of those $22 testers on Amazon and used it to find the bad one. The driver had failed.
I used the tester to check all my 13n10, 13n10L, and IRF520 mosfets. There were some definite differences in them. Not large, but noticeable.
C readings of the 13n10s were 712 to 718 pF, with Vt from 3.9 to 4.1v.
The ones that were removed read 712 and the lowest Vt at 3.7.
Attachments
You've got it!
The Vt is as you already know - the trigger voltage - so good to know the tester helps to develop a profile of the part BEFORE you try and put it in. At least you are aware of the variances - which is why I mentioned that "matching part".
The problem with "trigger" is many get confused with the performance of the part related to it's trigger - which helps you install the part, but does not give you a clue as to how the gate capacitance will affect how the part performs as the replacement.
Trigger is a relative voltage reference - that refers to how the device appears at that voltage it tested on as.
As you found out, the OEM part is/was a 3.7 which the replacement showed 4.1Vt is pretty close to 1/2V ABOVE the OEM one - so then you have to look at the power input to the Gate - as the "trigger level" - but not as a tuning or stability reference - which gate capacitance then becomes that figure in which to select from.
So I would have skipped the 4.1Vt and tried to locate another with Vt closer to the OEM - but that does not rule out the capacitance.
You may never find an exact match using Vt - but the paring - which was the question I see - relies on the balance of capacitance the OEM had to the one you're trying to put in there.
In Push-Pull or even single-ended supplies - Vt tells you it may need more power to stay linear, but Cg or the terminal capacitance it appears as, will determine the stability in there as a replacement.
If I may add a little more...
The question of the tester - the voltage is truly only a relative reference. There is the usual issues of how you zero out a device so sensitive that even the oil on your fingers and the flux in the wetted solder - or the oxides of a preservative to keep the part from rusting (reduction) - affects the trigger - due to the high impedance state the device you test - is working in - dirty tester claw or teeth - or even soldered on - doesn't matter - consistency in results does.
As you can see - even in the Datasheets, there is a variance in the dies results - so the burden of proof then becomes not what the device appears as - more in, how you test the device as well as the consistent results of that method of testing - provide. That gives you the predictability.
It's when you test a different Brand to another where you may find exceptional differences in results - but that may be from variables that if you had not tested using a method - you may never truly know how well the part will perform - until it is installed.
The Vt is as you already know - the trigger voltage - so good to know the tester helps to develop a profile of the part BEFORE you try and put it in. At least you are aware of the variances - which is why I mentioned that "matching part".
The problem with "trigger" is many get confused with the performance of the part related to it's trigger - which helps you install the part, but does not give you a clue as to how the gate capacitance will affect how the part performs as the replacement.
Trigger is a relative voltage reference - that refers to how the device appears at that voltage it tested on as.
As you found out, the OEM part is/was a 3.7 which the replacement showed 4.1Vt is pretty close to 1/2V ABOVE the OEM one - so then you have to look at the power input to the Gate - as the "trigger level" - but not as a tuning or stability reference - which gate capacitance then becomes that figure in which to select from.
So I would have skipped the 4.1Vt and tried to locate another with Vt closer to the OEM - but that does not rule out the capacitance.
You may never find an exact match using Vt - but the paring - which was the question I see - relies on the balance of capacitance the OEM had to the one you're trying to put in there.
In Push-Pull or even single-ended supplies - Vt tells you it may need more power to stay linear, but Cg or the terminal capacitance it appears as, will determine the stability in there as a replacement.
If I may add a little more...
The question of the tester - the voltage is truly only a relative reference. There is the usual issues of how you zero out a device so sensitive that even the oil on your fingers and the flux in the wetted solder - or the oxides of a preservative to keep the part from rusting (reduction) - affects the trigger - due to the high impedance state the device you test - is working in - dirty tester claw or teeth - or even soldered on - doesn't matter - consistency in results does.
As you can see - even in the Datasheets, there is a variance in the dies results - so the burden of proof then becomes not what the device appears as - more in, how you test the device as well as the consistent results of that method of testing - provide. That gives you the predictability.
It's when you test a different Brand to another where you may find exceptional differences in results - but that may be from variables that if you had not tested using a method - you may never truly know how well the part will perform - until it is installed.
- Just so you know - the GATES are usually 2 to 4V - so the tester is already maxing out at 4.1V to even know the device is there.
- That can send up a red flag to me on why is it maxed out as ABOVE a typical ON result in the Datasheets?
- Might have more to worry about - but then too, the Oxides of March...keep marching on
Last edited:
Uh, "Vee Tee Aitch"?
As in threshold voltage? Measure of the minimum DC gate-to-source potential that, if you will, "triggers" the channel (drain-to-source" current path to first turn on.
Dunno about "trigger", we always called it threshold. A rose by any other name?
And yes, each batch of this part we buy gets that measurement jotted onto its back so two parts with the closest match get paired in a repair job.
Probably matters more in the 4-transistor RCI linear where two pairs get wired directly parallel to one another.
73
As in threshold voltage? Measure of the minimum DC gate-to-source potential that, if you will, "triggers" the channel (drain-to-source" current path to first turn on.
Dunno about "trigger", we always called it threshold. A rose by any other name?
And yes, each batch of this part we buy gets that measurement jotted onto its back so two parts with the closest match get paired in a repair job.
Probably matters more in the 4-transistor RCI linear where two pairs get wired directly parallel to one another.
73
OK. I'm learning a lot here. I used to repair CBs back in the 80s, and have forgotten a lot in the last 30 years. Mosfets were nowhere near being used as a final device, let alone being used anywhere in a CB.
So, I have replaced the 3 pills with some with a higher trigger/threshold voltage. Is there reason to make an adjustment in the circuit to accommodate for that? Being as the radio is a Q5n, I don't believe there is an accurate schematic to go by.
Or, should I just put the OEM pills back in, as only the driver was bad?
Or, just leave it alone. It's working, but I have had some reports of slight audio clipping from local stations.
So, I have replaced the 3 pills with some with a higher trigger/threshold voltage. Is there reason to make an adjustment in the circuit to accommodate for that? Being as the radio is a Q5n, I don't believe there is an accurate schematic to go by.
Or, should I just put the OEM pills back in, as only the driver was bad?
Or, just leave it alone. It's working, but I have had some reports of slight audio clipping from local stations.
That Vt stuff were talking about is the test tells you "I got this when I hit it with that..."
So usually, the MOSFET is "tuned" by mA draw at idle to get the threshold right - that was what I updated that one post with a graphic about.
When the part tests with max Gate trigger - that does nothing to help you - that was the point of the post, you'll have to look more at what it "exhibited" when that tester "did this - to get that", moment.
Usually these things (Tester) uses 1MHz test frequency (it'd say so for RF purposes to even find the roll off/exhibited capacitance) so if the tester matches the datasheets - you should be ok. (Capacitance wise - not necessarily Vt)
But you still have to re-trim the Bias to fit the new parts to make them work together the best for longer life, cannot guarantee anything unless you do. So, for now, the thing works and seems ok, it can blow up again or you can run it for decades and not have any more issues.
So usually, the MOSFET is "tuned" by mA draw at idle to get the threshold right - that was what I updated that one post with a graphic about.
When the part tests with max Gate trigger - that does nothing to help you - that was the point of the post, you'll have to look more at what it "exhibited" when that tester "did this - to get that", moment.
Usually these things (Tester) uses 1MHz test frequency (it'd say so for RF purposes to even find the roll off/exhibited capacitance) so if the tester matches the datasheets - you should be ok. (Capacitance wise - not necessarily Vt)
But you still have to re-trim the Bias to fit the new parts to make them work together the best for longer life, cannot guarantee anything unless you do. So, for now, the thing works and seems ok, it can blow up again or you can run it for decades and not have any more issues.
I'm just recalling a short paragraph from Lou Franklin's book... He mentions how the bias in a class C amp is set up so that the conduction times of all the transistors are the same. He didn't get deep into the theory, unfortunately, but it has me wondering whether or not swapping in a part with a higher threshhold voltage will cause distortion (as the conduction time is changing significantly for that transistor).
Anyone able to elaborate on what Lou was getting at there on pg. 162?
I would have started a new thread for this but I think it applies to some tiny degree here.
Anyone able to elaborate on what Lou was getting at there on pg. 162?
I would have started a new thread for this but I think it applies to some tiny degree here.
I wonder if this has something to do with why some of the radio service manuals mentioned setting mosfet bias by voltage. Obviously that dont work with how fast the current changes with a tiny change in voltage, but now I wonder if they had conduction times in mind when writing that. Maybe both Vth and Idq are important?
I asked ChatGPT. According to that, the conduction angle doesn't need to match....
Nice Brandon!
i've been wanting to mess with chatGPT and CB radio questions but haven't found the time.
I would love to hear what it has to say about CB radio alignment and modification.
LC
I'm going to leave things alone. It's working. I appreciate all the responses.
@brandon7861 - the article did state something quite interesting that many people did not (even in older BJT days) catch.
That being the conduction angle (can also be considered % of wave) should and recommended and when you look at CB radios - a preferred means to generate the clean waveform input so the audio and other whatnots, can be applied to the signal.
So the Predriver sections or stages upstream from that stage of the amp - need to provide clean signal for input for driving that stage.
This usually indicates AB bias or even A to get things started.
That being the conduction angle (can also be considered % of wave) should and recommended and when you look at CB radios - a preferred means to generate the clean waveform input so the audio and other whatnots, can be applied to the signal.
So the Predriver sections or stages upstream from that stage of the amp - need to provide clean signal for input for driving that stage.
This usually indicates AB bias or even A to get things started.