The 8KHz band is thermal related? How can this conclusion be drawn? And at -90db? How audible can this be even if it were related with any kind of distortion?
I think a simulation that turns the drive signal on an off at maybe 0.25 seconds intervals and see what the rate-of-change of the resistance is, this way you can probably figure out the magnitude of possible modulation.
I think a simulation that turns the drive signal on an off at maybe 0.25 seconds intervals and see what the rate-of-change of the resistance is, this way you can probably figure out the magnitude of possible modulation.
nickmckinney said:
We certainly could do this with a multimeter too, *if* we are quick enough.
Maybe the better way is to follow my measurement setup and just adapting it to do several sine bursts .
Again we have the compression driver from above and again I inject a roughly 10W into the VC taken off the gap, in order to show effects taking place as clearly as possible.
An externally hosted image should be here but it was not working when we last tested it.
This measurement shows us how the energy of the 10kHz bursts melt down the current flow – and subsequently the efficiency of that speaker - as VC temperature rises
The upper trace shows the undistorted bursts.
An externally hosted image should be here but it was not working when we last tested it.
This "zoomed in" plot shows, how the bursts recover after each pause.
If we would extend our measurement until there is thermal equilibrium reached – we would even more clearly see that the raise of thermal distortion and the decay of TD are in fact symmetric – very close as predicted by my simulation.
This is independent of what material is the surrounding of the VC – meaning it does not matter for the qualitative behaviour basically – if the VC sticks in the gap or not.
nickmckinney said:
And also remember - once the motor is hot (at equilibrium to be precise) its exactly the stage where symmetric heating and cooling happen - to maintain that constant equilibrium temperature of the motor.
If you don't trust my prediction, I suggest you to try yourself – and hopefully report back.
Basically thermal behaviour isn't all that intuitive for most of us.
Maybe its also due to the fact, that heat flow is kind of a multidimensional "open" system - in contrast to electric circuits heat moves / spreads out – and this spreading out takes time.
One of the mental reasons that so many had ( have ) a problem with the *instantaneous* heat up of the VC wire.
The wire itself heats up *instantaneous* - yes, but even the enamel insulation around the VC wire is the first thermal mass that heats up "delayed" or better put – low pass filtered.
The same happens on the way of thermal flow with the next thermal mass – the VC former in case of conduction (or the inside of the whole motor in case of radiation) – and so on and so on and so on....
Until you reach the thermal sink of your room – your neighbourhood – whole earth – and in the end cause raise to the entropy of our universe.
gedlee said:
Me neither, Earl, me neither...
gedlee said:
I checked back and been told that subharmonics could be possible - even in time invariant systems – and I might have been looking at the wrong end.
May came back or not on NTD depending on measurement results. The thermal low pass in the range of the one digit Hz (estimated form the plots measured) makes measurements not really easy.
soongsc said:
Again – this part, especially Earl would like to verify, is more about quality (if and how it happens at all) than about quantity.
As for me, I'm already happy with the clarified results on the long time scale – though it would be tempting to know better about NTD as well .
Michael
mige0 said:
No speed needed, give it a real world try already.
I doubt the motor can get that hot as the VC would probably burn up long before. Think about what is needed to get an entire speaker motor to 400F which is the burning point of a typical voice coil. You will have localized hot spots in the motor closer to the voice coil.
Yes, I agree with Nick, I doubt that a driver ever gets to a steady state - well actually it would, called burned out. That state is pretty steady.
But I want to emphasize again that what is being talked about here are long term effects NOT short term ones like I'm interested.
Heck with a two channel sound card it would be easy to do a monitor of VC temp with signal. Its just a couple of FFTs and an average of the quotients - this is basically Re.
But I want to emphasize again that what is being talked about here are long term effects NOT short term ones like I'm interested.
Heck with a two channel sound card it would be easy to do a monitor of VC temp with signal. Its just a couple of FFTs and an average of the quotients - this is basically Re.
nickmckinney said:
You'd like to share your findings?
Could you possibly plot it on a time line?
nickmckinney said:
gedlee said:
A good one, Earl
But - again there is kind of a common mental hurdle involved IMO
Sure there is thermal equilibrium that can be reached (at least if we are patient enough to take an infinite time of observation
) !At what temperature exactly equilibrium is reached depends on the energy injection on one hand and on dissipation capability on the other hand.
So, for any moderate power injection there will be a thermal equilibrium at a certain motor temperature in a certain environment (without destroying the VC) - *if *that environment is assumed to keep basically constant temperature – as is our everyday experience of speakers (or do *you* have to change a burned VC after every piece of music?).
The final thermal reservoir of the universe is pretty huge, I guess – but if we consider it to be finite I would have to agree that thermal equilibrium in fact *can not* be achieved - even if we would observe a infinite time (unless we use the trick of air condition
).But its a rather academic discussion at what time universe reaches temperature of VC burn up...
gedlee said:
Again no, Earl – this is one of the reasons I have agreed on your concerns if my previously shown FFT holds as a proof for NTD.
The basic problem here is, that we inherently measure signals that are not steady. Or at least *I* did – not waiting for equilibrium happen.
For this kind of signals, FFT analysis is quite limited to tell us the "truth".
Michael
mige0 said:
I don't have any findings other than the experience of rebuilding an average 30-40 drivers per day at a large recone center. All makes, Ohm, Scanspeak, McIntosh, Altec, EV, Guass, WE, just to name a few. Also local car scene was really good at blowing drivers, and the local "rough neighborhood" scene had this infatuation with mulitple pro sound 10" drivers located where their backseat used to be. I have seen up to 16 EV10 drivers in a Chevy before SPL contests became mainstream.
First I started modifying current drivers mixing and matching parts trying to keep from seeing the same faces with the same blown equipment. Then I built a couple from scratch and had Rick at Ramsdell Audio magnetize them. They were the first Lambda drivers covered in black paint with edgewound coils from some forgotten JBL design. The rest is history when someone gave me the bright idea that I could sell more of them.
Lambda started in 1999 and ended in 2004 when I sold the majority of the hard goods to AE speakers. I spent 2004 until the early part of this year forgetting else everything I could speaker related. John from AE called and asked me to help him around April, and thats why I am back. As for papers on anything? Gone to the wind, I was surprised John was still here making them to be honest.
mige0 said:
We are not looking at the same things here. Something can be "short time" and still be "continuous". My point is that for there to be NTD there has to be somewhere in the thermal chain a time constant short enough to react at audio frequencies, i.e. about 20 ms or less. This can still happen continuously. Your looking at things happening in seconds - this is way below the audio bandwidth - like 1 Hz. This won't modulate an audio frequency in an audible way.
Loudspeaker heat management
I agree with what others have said on this thread, that there are two separate issues. One is the suggestion that rapid onset heating might occur within the time interval of a cycle in the passband. That would require a small thermal mass compared to the amount of energy put in. The second issue is basically an elevated local ambient temperature, something that causes a steady-state rise in voice coil temperature. This shifts the electro-mechanical parameters, and ultimately causes the voice coil adhesive to weaken and the driver to eventually fail completely.
When I was testing my cooling plug approach, I found the shift in local ambient temperature was significant. A driver with a cooling plug that wicks heat out of the center pole would survive indefinitely at power levels that would kill the exact same driver without a cooling plug, often in less than a couple hours. That's a big deal for prosound applications where continuous high power levels can destroy drivers during an event. Lots of sound crews end up reloading their cabinets with fresh drivers after shows, and the cooling plug approach stops all that. You can expect to see a 2x to 3x increase in power handling, which is good insurance.
One interesting phenomenon I saw was that at power levels approaching the maximum safe level, you could actually hear the reduction in SPL as a function of time. My testing method was to run a 40 cycle sine wave for 15 seconds on and 15 seconds off. After the driver warms up for about 10 minutes of these cycles at full power, you can hear the volume level drop through the 15 second on time. Audible compression set in that quickly.
From a cold start, the center pole temperature would rise to levels hot enough to boil water without a cooling plug, but only to about 120 degrees with the cooling plug installed. I later ran the tests much higher than the driver without a cooling plug would allow, as much as 220%, all without failure. Even at these power levels, the driver protected by a cooling plug was not nearly as hot as one without, running less than half the power. You can see the test results at the link below.
I agree with what others have said on this thread, that there are two separate issues. One is the suggestion that rapid onset heating might occur within the time interval of a cycle in the passband. That would require a small thermal mass compared to the amount of energy put in. The second issue is basically an elevated local ambient temperature, something that causes a steady-state rise in voice coil temperature. This shifts the electro-mechanical parameters, and ultimately causes the voice coil adhesive to weaken and the driver to eventually fail completely.
When I was testing my cooling plug approach, I found the shift in local ambient temperature was significant. A driver with a cooling plug that wicks heat out of the center pole would survive indefinitely at power levels that would kill the exact same driver without a cooling plug, often in less than a couple hours. That's a big deal for prosound applications where continuous high power levels can destroy drivers during an event. Lots of sound crews end up reloading their cabinets with fresh drivers after shows, and the cooling plug approach stops all that. You can expect to see a 2x to 3x increase in power handling, which is good insurance.
One interesting phenomenon I saw was that at power levels approaching the maximum safe level, you could actually hear the reduction in SPL as a function of time. My testing method was to run a 40 cycle sine wave for 15 seconds on and 15 seconds off. After the driver warms up for about 10 minutes of these cycles at full power, you can hear the volume level drop through the 15 second on time. Audible compression set in that quickly.
From a cold start, the center pole temperature would rise to levels hot enough to boil water without a cooling plug, but only to about 120 degrees with the cooling plug installed. I later ran the tests much higher than the driver without a cooling plug would allow, as much as 220%, all without failure. Even at these power levels, the driver protected by a cooling plug was not nearly as hot as one without, running less than half the power. You can see the test results at the link below.
gedlee said:
Ok - "short time" was somehow misleading, and I wanted to point out that NTD - if we can establish a measurement procedure - wouldn't happen at the beginning of the TD curve only.
As for audibility – I'd be more cautious on that – remember the "brain" frequencies at the low digit Hz for example ? But again, its pretty contra productive to discuss any audibility when the effect is neither known nor validated at all.
You could invest some of your talent here and support investigation with either measurements (the one you proposed) or do some further math analysis on the topic - if you are really interested - rather than taking an early seat at the balcony of the muppet show (which is OK as well
)
Michael
Re: Loudspeaker heat management
Thanks for providing some background here, Wayne.
What you outlined is pretty much what we see from my simus, measurements and the picture I try to draw about thermal effects in speakers.
The VC taken off the gap for most of my measurements so far - I didn't actually hear anything
Michaal
Wayne Parham said:
Thanks for providing some background here, Wayne.
What you outlined is pretty much what we see from my simus, measurements and the picture I try to draw about thermal effects in speakers.
The VC taken off the gap for most of my measurements so far - I didn't actually hear anything
Michaal
Re: Loudspeaker heat management
WAyne - exactly correct. The smaller the driver the lower the efficiency the higher the power needed and the smaller the thermal mass. Thus there does seem to be a reason to suspect that this effect can occur in a smaller system but not in the larger one. Maybe this is what we hear as better dynamics.
When I listen to a dome tweeter it just never seems to sound right at higher SPLs. At low levels they can be OK, but dynamically they just don't cut it.
I am very interested, I'm just swamped. I'm trying to establish a business and I'm doing this solo. That hasn't left me with much more time than to simply read E-mails and sometimes respond. Time permitting I will do all the tests that I suggested, but time is not plentiful at the moment.
Wayne Parham said:
WAyne - exactly correct. The smaller the driver the lower the efficiency the higher the power needed and the smaller the thermal mass. Thus there does seem to be a reason to suspect that this effect can occur in a smaller system but not in the larger one. Maybe this is what we hear as better dynamics.
When I listen to a dome tweeter it just never seems to sound right at higher SPLs. At low levels they can be OK, but dynamically they just don't cut it.
mige0 said:
You could invest some of your talent here and support investigation with either measurements (the one you proposed) or do some further math analysis on the topic - if you are really interested - rather than taking an early seat at the balcony of the muppet show (which is OK as well
I am very interested, I'm just swamped. I'm trying to establish a business and I'm doing this solo. That hasn't left me with much more time than to simply read E-mails and sometimes respond. Time permitting I will do all the tests that I suggested, but time is not plentiful at the moment.
I my tests, there was about 3dB drop over a 15 second period. It did not occur when the driver was cold, only after it had become hot and power was high. Of course, at high power levels, you will heat the driver in a relatively short period of time, not seconds but not hours either. It only takes 10 to 20 minutes, just a couple songs. After that the driver is hot and you could hear the gradual drop in output from the beginning to the end of the burst in my test cases.
Of course, this was a relatively slow effect. It was like a bias shift. The reduction in amplitude was about the rate of 0.5dB/second. After 5 or 6 seconds, the compression increase (and SPL decrease) was slower. So you would lose 3dB in about 5 seconds, and then a slower dropoff after that. This was at very high power levels near the maximum safe continuous level.
To me, this is mostly an issue about compression and longevity. But it is also true that during the SPL drop from compression, each half cycle is smaller than the one that preceeds it. The level is very small, but it is there. Naturally, this would introduce harmonic distortion. Not sure about the audibility of it, but it would seem logical that it should be there.
Of course, different drivers have different compression characteristics. They all have different time intervals before the onset of compression, the length of time before full compression is reached and the length of time before thermal failure. My tests were of a large and powerful woofer. Smaller drivers will have correspondingly smaller power levels and shorter time intervals. The better you can make heat management, the longer the intervals and the higher the power before thermal effects set in.
Of course, this was a relatively slow effect. It was like a bias shift. The reduction in amplitude was about the rate of 0.5dB/second. After 5 or 6 seconds, the compression increase (and SPL decrease) was slower. So you would lose 3dB in about 5 seconds, and then a slower dropoff after that. This was at very high power levels near the maximum safe continuous level.
To me, this is mostly an issue about compression and longevity. But it is also true that during the SPL drop from compression, each half cycle is smaller than the one that preceeds it. The level is very small, but it is there. Naturally, this would introduce harmonic distortion. Not sure about the audibility of it, but it would seem logical that it should be there.
Of course, different drivers have different compression characteristics. They all have different time intervals before the onset of compression, the length of time before full compression is reached and the length of time before thermal failure. My tests were of a large and powerful woofer. Smaller drivers will have correspondingly smaller power levels and shorter time intervals. The better you can make heat management, the longer the intervals and the higher the power before thermal effects set in.
In my system, I have a lot of driver surface area compared to the sound levels that I listen too (the woofers just don't move very much). Every time I touch the woofer magnets of my speakers, they feel cool to the touch - even after long listening sessions. Thus, I doubt that I am experiencing compression and distortion due to hot voice coils. Also, my drivers are in open baffles, which increases the heat loss compared to closed systems.
However, I am worried more about compression in my amps. The power transformers, and particularly the J-Fets, of my amps heat a lot more than the voice coils of my drivers (Hafler P7000).
What compression and distortion might I be experiencing from my amps? Particularly since open baffle designs are more demanding of amps than sealed systems.
Retsel
However, I am worried more about compression in my amps. The power transformers, and particularly the J-Fets, of my amps heat a lot more than the voice coils of my drivers (Hafler P7000).
What compression and distortion might I be experiencing from my amps? Particularly since open baffle designs are more demanding of amps than sealed systems.
Retsel
To start with putting our *qualitative* measurements so far into some meaningful specs – in order to create sort of standard for *quantifying* TD lets go back to what we already have seen:
and overlay this plot:
What we see here is how three time constants of 0,3sec (RED trace) 3sec (BLUE trace) and 30 sec (GREEN trace) look like on a time line
Even visually its quite obvious that the measured compression driver VC resistance (at 10W injection, taken off the gap) has a time constant around 3 seconds by simply comparing the two pix
We could do some more fine tuning to get the BLUE trace closer to the envelope of the measurement above, but in the end we possibly would only see that it never fits *exactly* because of the simplified model.
Even so it seems to be save to claim the roughly three seconds time constant in this case.
This would translate into a thermal low pass behaviour at around 0,05Hz (to correct my earlier "single digit" statement )
Thanks for the overview about Lambda / AE Speaker history .
Edgewound on the inside *and* the outside of a fibreglass former would possibly be my favourite for the upcoming 6.5" PRO
Are you involved in the design of this one too?
Would be pretty interesting to get some thermal figures in comparison to a one side wound dual layer VC on an aluminium former.
Its hard to do an estimation whether or not the better thermal conductivity of the aluminium former out weights the additional heatup of the former due to eddy currents.
My guess is, that aluminium former still should be better as we could possibly look at the trade off's from a mere power injection standpoint and this amount does not change.
But only measurements of the Thermal Distortion could clarify .
Measuring TD isn't really that hard to do – a procedure to get reliable results I have already described (and Wayne, Brian and other too have done investigation in a similar fashion) – the software I used is AUDACITY (freeware) and anything else needed is a decent shunt .
Might be that some time this might be something one wants to give a closer look – but for now and for me - its out of the scope of this thread / topic.
BTW – I wouldn't worry as Haflers are negative feedback amps AFAIK
Michael
An externally hosted image should be here but it was not working when we last tested it.
and overlay this plot:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
What we see here is how three time constants of 0,3sec (RED trace) 3sec (BLUE trace) and 30 sec (GREEN trace) look like on a time line
Even visually its quite obvious that the measured compression driver VC resistance (at 10W injection, taken off the gap) has a time constant around 3 seconds by simply comparing the two pix
We could do some more fine tuning to get the BLUE trace closer to the envelope of the measurement above, but in the end we possibly would only see that it never fits *exactly* because of the simplified model.
Even so it seems to be save to claim the roughly three seconds time constant in this case.
This would translate into a thermal low pass behaviour at around 0,05Hz (to correct my earlier "single digit" statement
)nickmckinney said:
Thanks for the overview about Lambda / AE Speaker history .
Edgewound on the inside *and* the outside of a fibreglass former would possibly be my favourite for the upcoming 6.5" PRO
Are you involved in the design of this one too?
Would be pretty interesting to get some thermal figures in comparison to a one side wound dual layer VC on an aluminium former.
Its hard to do an estimation whether or not the better thermal conductivity of the aluminium former out weights the additional heatup of the former due to eddy currents.
My guess is, that aluminium former still should be better as we could possibly look at the trade off's from a mere power injection standpoint and this amount does not change.
But only measurements of the Thermal Distortion could clarify .
Measuring TD isn't really that hard to do – a procedure to get reliable results I have already described (and Wayne, Brian and other too have done investigation in a similar fashion) – the software I used is AUDACITY (freeware) and anything else needed is a decent shunt .
Retsel said:
Might be that some time this might be something one wants to give a closer look – but for now and for me - its out of the scope of this thread / topic.
BTW – I wouldn't worry as Haflers are negative feedback amps AFAIK
Michael
mige0 said:
Don't hold your breath as I am not a fan of inside/outside coils at the moment. They come with other problems and little to zero gain over our regular 4 layer flatwires. You will probably also never see an edgewound coil from us as they are inferior to 4 layer flatwires since they need 90 degree aluminum solder "blob" joints which few people can do well.
I am involved on all new AE drivers since signing on this spring.
Aluminum former eddy currents I don't worry about, the formers are not only very thin in comparison they do not form a complete shorted loop. They are so thin actually that we use a layer of Kapton above the voice coil wire to help strengthen the former where the cone and spider attach.
The main issue with aluminum formers is I don't like to use them in most underhung designs as the RMS could vary depending on xmax. Most manufacturers don't use them because they have such pathetic heatsinking in their motors the former carries all the voice coil heat directly into the cone and spider. I am sure you have heard of speaker cones that have caught on fire and glue bonds that break.
We are going a few different ways with the 6.5 driver, keep your eyes open.
To start with putting our *qualitative* measurements so far into some meaningful specs – in order to create sort of standard for *quantifying* TD lets go back to what we already have seen:
and overlay this plot:
What we see here is how three time constants of 0,3sec (RED trace) 3sec (BLUE trace) and 30 sec (GREEN trace) look like on a time line
Even visually its quite obvious that the measured compression driver VC resistance (at 10W injection, taken off the gap) has a time constant around 3 seconds by simply comparing the two pix
We could do some more fine tuning to get the BLUE trace closer to the envelope of the measurement above, but in the end we possibly would only see that it never fits *exactly* because of the simplified model.
Even so, it seems to be save to claim the roughly three seconds time constant in this case.
This would translate into a thermal low pass behaviour at around 0,05Hz (to correct my earlier "single digit" statement )
Thanks for the overview about Lambda / AE Speaker history .
Edgewound on the inside *and* the outside of a fibreglass former would possibly be my favourite for the upcoming 6.5" PRO
Are you involved in the design of this one too?
Would be pretty interesting to get some thermal figures in comparison to a one side wound dual layer VC on an aluminium former.
Its hard to do an estimation whether or not the better thermal conductivity of the aluminium former out weights the additional heatup of the former due to eddy currents.
My guess is, that an aluminium former still should be better as we could possibly look at the trade off's from a mere power injection standpoint and this amount does not change.
But only measurements of the Thermal Distortion will tell us definitely .
Measuring TD isn't really that hard to do – a procedure to get reliable results I have already described (and Wayne, Brian and others too have done investigation in a similar fashion) – the software I used is AUDACITY (freeware) and anything else needed is a decent shunt .
To do some further steps towards "developing on a TD point of view" lets have a look at the next simu's
What we see here is how a change of thermal mass of the VC (one order of magnitude in either way ) affects resistance change
Uhh – obviously quite a difference !
Same as above but shown on a different timescale.
Different mass obviously changes time constant of the VC – *if* we could freely select thermal mass.
No surprise – R*C= time constant.
But there is kind of surprise for short time energy injection
For low thermal mass VC's resistance melts down pretty quick and recovers pretty quick - but adds very *heavily* in short time amplitude.
All traces are tending towards the same figure for the mean value - as a result of equal power injection for all three traces.
Ok *now* its kind of intuitively as well.
There is a conclusion that can be drawn not obvious in the first place.
Good cooling always is a good thing – as we already know !
On top of that, adding thermal mass to the VC is a good thing too.
Something to think about
Aluminium for example has roughly double the thermal capacity of copper.
I'm no specialist for this kind of trade off's in magnetic circuits (somehow who can do a quick calculaion?) as there is also lower specific mass and slightly higher resistance compared to copper – but my guess is that there might be a net benefit in increase of sensitivity due to lower mass *and* possibly a slight increase of effective thermal mass too.
Both good things with regard to Thermal Distortion – as long as there are no alternative VC wire materials with *much* lower temp-co .available.
Might be that some time this might be something one wants to give a closer look – but for now and for me - its out of the scope of this thread / topic.
BTW – I wouldn't be concerned at all, as Haflers are negative feedback amps AFAIK
Michael
PS
Sorry for messing up the record
This posting should have been at #58
An externally hosted image should be here but it was not working when we last tested it.
and overlay this plot:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
What we see here is how three time constants of 0,3sec (RED trace) 3sec (BLUE trace) and 30 sec (GREEN trace) look like on a time line
Even visually its quite obvious that the measured compression driver VC resistance (at 10W injection, taken off the gap) has a time constant around 3 seconds by simply comparing the two pix
We could do some more fine tuning to get the BLUE trace closer to the envelope of the measurement above, but in the end we possibly would only see that it never fits *exactly* because of the simplified model.
Even so, it seems to be save to claim the roughly three seconds time constant in this case.
This would translate into a thermal low pass behaviour at around 0,05Hz (to correct my earlier "single digit" statement
)nickmckinney said:
Thanks for the overview about Lambda / AE Speaker history .
Edgewound on the inside *and* the outside of a fibreglass former would possibly be my favourite for the upcoming 6.5" PRO
Are you involved in the design of this one too?
Would be pretty interesting to get some thermal figures in comparison to a one side wound dual layer VC on an aluminium former.
Its hard to do an estimation whether or not the better thermal conductivity of the aluminium former out weights the additional heatup of the former due to eddy currents.
My guess is, that an aluminium former still should be better as we could possibly look at the trade off's from a mere power injection standpoint and this amount does not change.
But only measurements of the Thermal Distortion will tell us definitely .
Measuring TD isn't really that hard to do – a procedure to get reliable results I have already described (and Wayne, Brian and others too have done investigation in a similar fashion) – the software I used is AUDACITY (freeware) and anything else needed is a decent shunt .
To do some further steps towards "developing on a TD point of view" lets have a look at the next simu's
An externally hosted image should be here but it was not working when we last tested it.
What we see here is how a change of thermal mass of the VC (one order of magnitude in either way
) affects resistance changeUhh – obviously quite a difference !
An externally hosted image should be here but it was not working when we last tested it.
Same as above but shown on a different timescale.
Different mass obviously changes time constant of the VC – *if* we could freely select thermal mass.
No surprise – R*C= time constant.
But there is kind of surprise for short time energy injection
For low thermal mass VC's resistance melts down pretty quick and recovers pretty quick - but adds very *heavily* in short time amplitude.
All traces are tending towards the same figure for the mean value - as a result of equal power injection for all three traces.
Ok *now* its kind of intuitively as well.
There is a conclusion that can be drawn not obvious in the first place.
Good cooling always is a good thing – as we already know !
On top of that, adding thermal mass to the VC is a good thing too.
Something to think about
Aluminium for example has roughly double the thermal capacity of copper.
I'm no specialist for this kind of trade off's in magnetic circuits (somehow who can do a quick calculaion?) as there is also lower specific mass and slightly higher resistance compared to copper – but my guess is that there might be a net benefit in increase of sensitivity due to lower mass *and* possibly a slight increase of effective thermal mass too.
Both good things with regard to Thermal Distortion – as long as there are no alternative VC wire materials with *much* lower temp-co .available.
Retsel said:
Might be that some time this might be something one wants to give a closer look – but for now and for me - its out of the scope of this thread / topic.
BTW – I wouldn't be concerned at all, as Haflers are negative feedback amps AFAIK
Michael
PS
Sorry for messing up the record
This posting should have been at #58
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.