yes. NTC's in parallel is NO NO. there will NOT be the same impedance between the 2. and then the one with the lower impedance will start a current hog wich will make it hog more and more current as the impedance goes down. in other Words. the other parallel NTC will not be effective and your current capasity is halfed. and impedance will be twice the intended impedance.
Hi AudioSan,
If you did not have a relay to bypass the NTCs then what you say might happen and be a problem but once the relay pulls in and removes them from the current path there can be no problem. Paralleling NTCs is used in professional designs, where do you think I get the circuit, and the softstart boards you can buy are made that way also. I have been using the design for a long time with no problems. With lower power amps using one NTC and leaving it in the primary circuit is fine but for higher power amps I do not like to restrict the primary current after you get passed the turn on in-rush current.
Have a good one,
John
If you did not have a relay to bypass the NTCs then what you say might happen and be a problem but once the relay pulls in and removes them from the current path there can be no problem. Paralleling NTCs is used in professional designs, where do you think I get the circuit, and the softstart boards you can buy are made that way also. I have been using the design for a long time with no problems. With lower power amps using one NTC and leaving it in the primary circuit is fine but for higher power amps I do not like to restrict the primary current after you get passed the turn on in-rush current.
Have a good one,
John
this problem start the moment you puch the on button!
and yes. China E-bay softstarts use parallel NTC's. but china E-bay stuff is often UNSAFE.
NTC's should NEVER EVER be Connected in parallel. EEEEVER. its pointless at best. its a fire hazard at worst.
and yes. China E-bay softstarts use parallel NTC's. but china E-bay stuff is often UNSAFE.
NTC's should NEVER EVER be Connected in parallel. EEEEVER. its pointless at best. its a fire hazard at worst.
Last edited:
JohmH
listen !
NTC's in series are OK,
NTCs in parallel are not OK.
How many times does it need to be written for you?
listen !
NTC's in series are OK,
NTCs in parallel are not OK.
How many times does it need to be written for you?
Please provide example of such professional designs that I know which products to avoid.
Do NOT use parallel NTC
__________________
regards Andrew T.
Right now, to me, that is your opinion. What data/source do you have to justify that position. No word of mouth please. I don't take anything at face value, I need data.
Thanks,
John
__________________
regards Andrew T.
Right now, to me, that is your opinion. What data/source do you have to justify that position. No word of mouth please. I don't take anything at face value, I need data.
Thanks,
John
Do the math. 2x10ohm Negative Temperature Coefficient Thermistors with say 20% tollerance connected in parallel will not have exact the same resistance at the same temperature.
Electrons will ALLWAYS take the path With the least resistance. more current = lower resistance. Lower resistance = more current.
one thermistor is 9.8ohm. the other is 10.2ohm at 20c. witch thermistor will the current flow through?
the 9.8ohm. what happens when the current starts to flow??? the 9.8ohm heats up. what happens when that one heats up? it drops the resistance all the way Down to 0.1ohm. while the other one still is 20c and 10.2ohm.
Electrons will ALLWAYS take the path With the least resistance. more current = lower resistance. Lower resistance = more current.
one thermistor is 9.8ohm. the other is 10.2ohm at 20c. witch thermistor will the current flow through?
the 9.8ohm. what happens when the current starts to flow??? the 9.8ohm heats up. what happens when that one heats up? it drops the resistance all the way Down to 0.1ohm. while the other one still is 20c and 10.2ohm.
Last edited:
Hi AudioSan,
I fully understand the argument and agree with the possibility. In your arguement the lower resistance device will flow more current and heat up faster but the other device will flow current according to its resistance and still heat up, if more slowly. But let me qualify with so long as each device, NTC, has sufficent rating to withstand the initial surge than the device will not fail. I will search for supporting evidence of your arguement when I have time but I have experimented with this circuit for some time and it has withstood the tests. Now, with the new higher current power supply with massive storage the devices may not survive testing but then I will just select a device capable of higher currents and test with that.
Thanks,
John
I fully understand the argument and agree with the possibility. In your arguement the lower resistance device will flow more current and heat up faster but the other device will flow current according to its resistance and still heat up, if more slowly. But let me qualify with so long as each device, NTC, has sufficent rating to withstand the initial surge than the device will not fail. I will search for supporting evidence of your arguement when I have time but I have experimented with this circuit for some time and it has withstood the tests. Now, with the new higher current power supply with massive storage the devices may not survive testing but then I will just select a device capable of higher currents and test with that.
Thanks,
John
the difference in temperature between those 2 devices will increase exponentially very fast. and resistance decrease. therfor one of those devices wil allmost not heat up at all. and the reminding device must take all the current.
I'll give you Epcos TDK one of the many manufacturer's of NTC thermistors warning instead.
http://www.epcos.com/blob/528070/download/4/pdf-inrushcurrentlimiting-an2.pdf.
Page 3, bottom right of the page.
So basically you're going to choose an NTC that will function by itself.
Then what's the other one for?
I have an unrelated question.. two actually.
1) When mounting two boards on one 10x10 heatsink, do you mount so that you have transistors on the bottom of both boards or transistors in the middle of the heatsink? ( I think middle)
2) When using 4 pairs of outputs, are 2 pairs of input jFETS recommended? ( I think so. )
Thank you!
1) When mounting two boards on one 10x10 heatsink, do you mount so that you have transistors on the bottom of both boards or transistors in the middle of the heatsink? ( I think middle)
2) When using 4 pairs of outputs, are 2 pairs of input jFETS recommended? ( I think so. )
Thank you!
are you sure the 10x10" sink will do the job? is it by any chance heatsinkUSA's 10.080" profiles? i would cut the sinks in half. so i got 2 pc 10x5" sinks. that will be much more efficient.
Yes, it is the 10.080 profile. I won't be cutting them in half. If you put all the transistors at the middle, then the heat flow will go from middle to the outsides. The high temp presented by the other set of transistors will be a "virtual cut".
It will be enough heatsink for what I want to accomplish. I do not want to pump out 150 watts RMS class A. half of that is good enough.
It will be enough heatsink for what I want to accomplish. I do not want to pump out 150 watts RMS class A. half of that is good enough.
I think we are in agreement on this, one device could hog the current. And if each device can handle the load than everything is ok. If not and there is a failure then I will change over to one very large thermistor. I have found some in the catalog that are not too expensive.
Thanks,
John
Thanks,
John
The design was a direct copy and has worked fine, I may not be stressing it currently. But after reading the articles I am convinced, Thanks for the references, I will change the circuits to one thermistor of appropriate rating.
In a week or so the large transformers will be delivered and I will see if it can charge bank of capacitors without sucking out the fuse.
Thanks,
John
the lenght of the sinks will make it a little un-efficient. if you cut them in half they can acctualy handle allmost twice the power. you will need te spread out the heat. but then it will be un-even temperature between the output devices.
10.080x10" will give you about 0.4C/W pr sink(according to the specs). 75W class A at 8ohm is 2.1A bias. and you need just under +/-40V rails. thats 168W dissipation. With 25c ambient, you need about 0.15C/W pr ch too stay under 55c at the sinks.
if you are determined to og with the plan, i would mount the bords with the outputs a little bit apart from each other.
10.080x10" will give you about 0.4C/W pr sink(according to the specs). 75W class A at 8ohm is 2.1A bias. and you need just under +/-40V rails. thats 168W dissipation. With 25c ambient, you need about 0.15C/W pr ch too stay under 55c at the sinks.
if you are determined to og with the plan, i would mount the bords with the outputs a little bit apart from each other.