IF you fully heat them up and require them to coll down to the fully cold temperature.
If all you do is pass a current that subsides from ~90% of max rating to ~5% of max rating over a period of around 10 to 15 seconds, the device is not going to get anywhere near it's maximum temperature.
I have "seen" the effect of a Power NTC in the primary circuit of a Krell Klone PSU and what it did to the maximum output of the Klone. It all happened in less than 2seconds.
If all you do is pass a current that subsides from ~90% of max rating to ~5% of max rating over a period of around 10 to 15 seconds, the device is not going to get anywhere near it's maximum temperature.
I have "seen" the effect of a Power NTC in the primary circuit of a Krell Klone PSU and what it did to the maximum output of the Klone. It all happened in less than 2seconds.
No, you are not drawing max current into the smoothing caps for all that time (40s).
The caps are substantially charged in a few time constants.
15r and 10mF gives RC = 150ms.
5 time constants is 750ms.
If the NTC resistance is dropping below that start value of 15r, then 5*RC is less than 750ms
At the lower currents after the first couple of RC periods the NTC starts to cool.
It's resistance value to heading back towards it's cold value.
If you have a BIG ClassA bias current then the NTC will remain warm after start up. That warm resistance will not be the hot resistance. But the voltage drop of warm resistance* bias current can be fairly close to the voltage drop of max charging current * hot resistance.
i.e. you can be dropping a volt or so across the NTC. That is equivalent to having a high source impedance to your transformer/PSU.
High source impedance can significantly reduce amplifier performance !!!!!
You must switch out soft start NTC and switch out slow charge NTC, just as one would for fixed resistors. It's only the delay period that is different for fixed vs NTC.
The caps are substantially charged in a few time constants.
15r and 10mF gives RC = 150ms.
5 time constants is 750ms.
If the NTC resistance is dropping below that start value of 15r, then 5*RC is less than 750ms
At the lower currents after the first couple of RC periods the NTC starts to cool.
It's resistance value to heading back towards it's cold value.
If you have a BIG ClassA bias current then the NTC will remain warm after start up. That warm resistance will not be the hot resistance. But the voltage drop of warm resistance* bias current can be fairly close to the voltage drop of max charging current * hot resistance.
i.e. you can be dropping a volt or so across the NTC. That is equivalent to having a high source impedance to your transformer/PSU.
High source impedance can significantly reduce amplifier performance !!!!!
You must switch out soft start NTC and switch out slow charge NTC, just as one would for fixed resistors. It's only the delay period that is different for fixed vs NTC.
Last edited:
Please find attached the simulation I have been using. I do not pretend to be an LTspice guru, but this is what I have come up with. There are 4 CRCs, one per rail for each channel ( probably wrong terminology, but wth)
It uses the epcos thermistor libraries. The file ntc.txt is to be renamed ntc.asy , as diyaudio does not allow asy files to be attached.
The simulation runs for 40 seconds, then the thermistor bypasses are switched on and the sim continues to run. You can examine the current spikes in the NC switches and their bypass switches, as well as the power through one of the seconday side inductors that makes up the transformer.
Some switches are marked NC others NO. That shows their operation.
The internal resistance of the transformer consisting of the coupled inductors, was chosen to match the inrush of a similar transformer modelled on Duncan amp tools. The inductances were chosen to provide the correct secondary voltages that will be in my amp.
The switch delay time can be changed by modifying the delay param.
The load resistors are intended to model the amplifier bias a idle. Constant current sources were not used.
If there are any problems with this sim, please let me know.
It uses the epcos thermistor libraries. The file ntc.txt is to be renamed ntc.asy , as diyaudio does not allow asy files to be attached.
The simulation runs for 40 seconds, then the thermistor bypasses are switched on and the sim continues to run. You can examine the current spikes in the NC switches and their bypass switches, as well as the power through one of the seconday side inductors that makes up the transformer.
Some switches are marked NC others NO. That shows their operation.
The internal resistance of the transformer consisting of the coupled inductors, was chosen to match the inrush of a similar transformer modelled on Duncan amp tools. The inductances were chosen to provide the correct secondary voltages that will be in my amp.
The switch delay time can be changed by modifying the delay param.
The load resistors are intended to model the amplifier bias a idle. Constant current sources were not used.
If there are any problems with this sim, please let me know.
Last edited:
open the .asc file in LTspice (I assume you have it installed)
it is on the schematic
right-click on the .inc statement to edit it - remove the file path
.inc ntc_20130313.lib
also make sure ntc.txt (renamed to ntc.asy) is in the same folder (this is the Spice symbol for the ntc)
(it's "bad form" to share such work with dedicated file paths that only work on the originator's computer but not a biggie to fix - better to share a folder and have statements such as .inc without file paths and the .lib .txt files in the folder)
it is on the schematic
right-click on the .inc statement to edit it - remove the file path
.inc ntc_20130313.lib
also make sure ntc.txt (renamed to ntc.asy) is in the same folder (this is the Spice symbol for the ntc)
(it's "bad form" to share such work with dedicated file paths that only work on the originator's computer but not a biggie to fix - better to share a folder and have statements such as .inc without file paths and the .lib .txt files in the folder)
Last edited:
Yes, I have LTspice installed.
It reported
I have scanned through the 8 menus at the top of LTspice, but I can't see .inc
It reported
I have scanned through the 8 menus at the top of LTspice, but I can't see .inc
translate this into english please.
Last edited:
Not in the menus - on the schematic. Look down the left side.
It would have been less problematic for BigE to have shared a folder of files rather than individual files and have the .inc statement just be as I edited it above. At the moment the .inc has a specific file path that only works on his computer.
Andrew, create a folder - any folder, anywhere. Place the 3 files in that folder. Be sure to rename ntc.txt to ntc.asy as instructed by BigE (for reasons he gave). Edit the .inc statement on the schematic to
.inc ntc_20130313.txt
so that it matches the file you put in your folder. And you should be done.
ntc_20130313.txt is to be included with an .inc statement on the schematic. ntc.txt should be placed in the same directory as the .asc file and renamed to to ntc.asy.
Sorry for the confusion.
Next time, I will zip the files and use a portable .inc statement. I am new at this stuff.....
Sorry for the confusion.
Next time, I will zip the files and use a portable .inc statement. I am new at this stuff.....
Last edited: