Is there any way to avoid the high power dissapation in cap charging resistor R5?

Thread Starter

mike__MecE

Joined Feb 20, 2022
69
Hi, I endeavor to avoid the high wattages seen by R5 and the current loss through R5 (that is not going through my solenoid). On the other hand, I do not want to increase R5 because then I cannot cycle the solenoid according to the preferred frequency (10-20 Hz). I was thinking I could add a p-channel depletion-mode MOSFET above R5 that would be signaled at the same time as the input to the other 2 MOSFETs and prevent current from running through the cap charging leg. Does this idea have any promise? I can't figure out how to simulate it. Specifically, I can't figure out how to bring the depletion p-most gate to 0v at the same time the 3.3v from the micro is triggering the N-channel MOSFETs. Any help/suggestions would be appreciated.

Note: The LM317 is there because I am trying to limit the average power dissipation in the solenoid to 9W, regardless of the switching frequency.

Ref- https://archive.ednasia.com/www.ednasia.com/ART_8800523816_1000007_AN_dac7d6d3.HTM
(Image of example circuit schematic from the same publication attached)
 

Attachments

drjohsmith

Joined Dec 13, 2021
852
an alternative
is to charge the capacitor with a "constant current"
then use that power to move the solenoid,
a LM317 HV makes a great constant current source,
 

crutschow

Joined Mar 14, 2008
34,281
Since D1 and D2 are back-to-back Schottky diodes, they do not do anything in your circuit expect D1 will likely break down when the inductive spike exceeds its voltage rating (which can zap it).

What is the maximum switching frequency?
Is the pulse width a constant, independent of frequency?

If you tell us the exact purpose of the circuit, we can better help you.

Edit: OK, I see what R5 and C1 are supposed to do.
 
Last edited:

Thread Starter

mike__MecE

Joined Feb 20, 2022
69
Since D1 and D2 are back-to-back Schottky diodes, they do not do anything in your circuit expect D1 will likely break down when the inductive spike exceeds its voltage rating (which can zap it).

What is the maximum switching frequency?
Is the pulse width a constant, independent of frequency?

If you tell us the exact purpose of the circuit, we can better help you.

Edit: OK, I see what R5 and C1 are supposed to do.
This is a solenoid for actuating musical percussion.

Oops, there was supposed to be a Zener in there; fixed.

Frequency is 20 Hz max, looking for good performance at 16Hz and great performance at <10Hz (whereby performance I mean delivery of >.75J to the solenoid by each 10ms pulse (avg. 100W for the 10ms pulse)

I also want to actuate the solenoid as fast as possible, so I like that in my circuit the current spikes to near 8 amps and then declines as the caps discharge.

The pulse length will be adjustable from about 5-15ms, and I want to keep solenoid average power dissipation below 9W.
 

crutschow

Joined Mar 14, 2008
34,281
I am trying to limit the average power dissipation in the solenoid to 9W, regardless of the switching frequency.
The average power dissipation in R1 (which I assume is the solenoid resistance) is over 15W average at 10pps in your circuit.
If you reduce the input pulse width to 3ms, the average power drops to 8W, but that may not be sufficient time to actuate the solenoid.

Note that the average power is inversely proportional the the pulse frequency also.
 
Last edited:

Thread Starter

mike__MecE

Joined Feb 20, 2022
69
The average power dissipation in R1 (which I assume is the solenoid resistance) is over 15W average at 10pps in your circuit.
If you reduce the input pulse width to 3ms, the average power drops to 8W, but that may not be sufficient time to actuate the solenoid.

Note that the average power is inversely proportional the the pulse frequency also.
I am seeing approx. 9 watts or less when I sim different pulse widths and frequencies.
*Noting the circuit has changed since the original post, based on feedback from drjohsmith
1648773816664.png
 

Attachments

Thread Starter

mike__MecE

Joined Feb 20, 2022
69
If you want a simpler solution have a look at Toshiba TB67H450.
https://toshiba.semicon-storage.com/info/docget.jsp?did=65346&prodName=TB67H450FNG
It will drive a solenoid up to 3.5A @50V, and has a current limit input which makes it do PWM to reduce the average current without dissipating any power.
Thank you for sharing this. The Toshiba chip looks capable. I looked at a few motor and solenoid drivers and gate drivers. First, I would like a circuit that won't burn out the solenoid if the switch fails, and the circuit I posted is limited to the 100% duty cycle solenoid wattage by the LM317 if the switching MOSFET fails closed. Second, the circuit I posted applies nearly 8 amps for 1-3 ms and then the amperage drops to 1-2 amps, which delivers a lot of power quickly, for fast-as-possible actuation. Then the current drops which has the added benefit of less kickback when the switch turns off. I couldn't find a driver that could provide the amperage I want. Looked at the TI line including the DRV8876 and some others. I also don't want a programmable IC nor do I really want to hook up SPI for my gate driver and monitor current sense, etc. But maybe later in the project evolution, those will become necessary and I will defeat this comment.
 

Thread Starter

mike__MecE

Joined Feb 20, 2022
69
It will be higher at you maximum pulse rate of 10pps.

The average current draw from the 24V supply is nearly an amp.
Will that be a problem?
The current limiters seem to be doing their job and pulsing as fast as 20pps (50ms period) is <6W average.
Yes, I'm going to use a 24V 6A adapter and run a few of these circuits in parallel, maybe 4 .

1648781109288.png
 

Sensacell

Joined Jun 19, 2012
3,432
Note that the Max Vgs rating on M1 is only +- 16 volts.

When M2 turns on, the Source of M1 goes to minus 20-ish volts, the gate is at +10, the total Vgs could be 30V!
You need to clamp the Vgs of M1 to prevent damage.
 

drjohsmith

Joined Dec 13, 2021
852
The current limiters seem to be doing their job and pulsing as fast as 20pps (50ms period) is <6W average.
Yes, I'm going to use a 24V 6A adapter and run a few of these circuits in parallel, maybe 4 .

View attachment 264020
The Key to solenoids and power dissipation,
is to remember that the current needed initially is typically many powers higher than the hold current

In a drinks dispenser for instance, we needed 100A to move a big solenoid actuator
but hold current was less than 1 amp
( You might have heard the big "clang" at the start of a drinks dispensing on some machines )

The "answer" was we charged a big capacitor at a constant / max current of 1 Amp
And then short the capacitor across the solenoid to "fire it"
the capacitor moved the actuator, the constant current help the solenoid,
and then charged the capacitor back up when the solenoid was released.
( thats why its slow to enter what you want to drink, gives capacitor time to re charge )
 

Thread Starter

mike__MecE

Joined Feb 20, 2022
69
The Key to solenoids and power dissipation,
is to remember that the current needed initially is typically many powers higher than the hold current

In a drinks dispenser for instance, we needed 100A to move a big solenoid actuator
but hold current was less than 1 amp
( You might have heard the big "clang" at the start of a drinks dispensing on some machines )

The "answer" was we charged a big capacitor at a constant / max current of 1 Amp
And then short the capacitor across the solenoid to "fire it"
the capacitor moved the actuator, the constant current help the solenoid,
and then charged the capacitor back up when the solenoid was released.
( thats why its slow to enter what you want to drink, gives capacitor time to re charge )
Thanks. I would love to see the circuit or at least the solenoid, switch, and capacitor specs.
 

drjohsmith

Joined Dec 13, 2021
852
Thanks. I would love to see the circuit or at least the solenoid, switch, and capacitor specs.
It was a long time ago
The constant current charger was a LM317
We just had a "fet" as a switch to connect the capacitor to the Solenoid,
very simple,

We did a good few experiments with the coil to find out what currents were needed,
 

Thread Starter

mike__MecE

Joined Feb 20, 2022
69
Note that the Max Vgs rating on M1 is only +- 16 volts.

When M2 turns on, the Source of M1 goes to minus 20-ish volts, the gate is at +10, the total Vgs could be 30V!
You need to clamp the Vgs of M1 to prevent damage.
Thank you! Before and after (blue gate M1, red source M1, green current through solenoid)
Unfortunately, the peak current in my solenoid drops a bit but I assume that's because the excessive Vgs was opening M1 fast.
Clamped at 12V for now. Thanks again.
1648815432709.png

clamped.PNG
 
Top