Is it possible to use the 3.3v output logic voltage of arduino and the drive voltage of 5v of the shift register?
1 Like
It would have been better if you had asked this in your previous post where that 595 is used to drive mosfets.
http://forum.arduino.cc/index.php?topic=432953.0
A 3.3volt Arduino can't control 5volt HC logic directly.
The 74HC595 inputs would need at least ~3.5volt to register a high.
You could power the 595 from 3.3volt. Then the 595 will work.
But then you could have saturation problems with the mosfets.
The FQP50N06 was already a bad choice on 5volt, and probably won't work on 3.3volt.
Leo..
Wawa,
I just connected the 3.3v voltage to the 74hc595.
It works, but as you say, the FQP50N06 has a voltage drop.
The reason for using the 3.3v voltage is to use the high clock speed of the Arduino.
But I did not find the 84MHz Arduino, which works on 5v.
Thanks for the answer
From your previous post...
"I want to control many electromagnets."
That's doesn't require a fast Arduino.
Why do you think you need a faster Arduino.
Leo..
FQFP50N06 is inappropriate. It requires 10v on the gate to turn on completely, and has been "not recommended for new designs" (ie, obsolete) since 2009.
You should use a modern logic level MOSFET. An IRF3708pbf will switch up to 30v with just 3.3v on the gate (albeit with somewhat higher Rds(on) ), and is available in SMD and through-hole packages. For lighter loads, smaller MOSFETs are available, though if you want them to work with 3.3v on the gate, you won't find any in through-hole packages (the one I mentioned above is the only acceptable MOSFET available in a through-hole package that turns on fully with 3.3v on the gate - it's a fairly beefy MOSFET). There are lots of SOT-23 mosfets good for an amp or few that achieve mind-bogglingly low on-state resistance. If you need it to hold off a higher voltage, you'll need to use a 4.5v logic level MOSFET I think (I looked recently and couldn't find any appealing candidates that would work with 3.3v on the gate and >30Vds), and then you're back to the issue with having to shift the levels somehow.
Is it possible to use the 3.3v output logic voltage of arduino and the drive voltage of 5v of the shift register?
Wawa,
I want to use the video source (24x24 pixels) to operate 576 electromagnets.
Also, I want to adjust the magnetic force by PWM.
DrAzzy,
I saw the Vgs graph in the datasheet and it helped me to understand.
But sot-23 is very difficult to solder.
The IRF3708pbf that you told me is unfortunately expensive. (I need 576ea ...)
I plan to make only mosfet with the SMT, or plan to use a lot of 5v logic's Arduino to complement the clock speed.
Thank you very much.
dlloyd,
Can the SN74AHCT595 be used with different voltages at the same time? Sorry, I look to the data sheet do not understand.
Didn't think of using a HCT.
The SN74AHCT595 has different (lower) switch points. 0.8volt and 2volt on a 5volt supply.
Should work on 3.3volt logic.
This is the first time you have mentioned 567 solenoids. I think they were 12volt/1.3Amp each.
Wow.
12volt/737Amp
8.85kW
I hope you know what you're doing.
A large number of 595s (71) also changes driving them.
A 3.3volt MCU port probably hasn't got the fanout power for that.
And 595s don't do PWM natively.
You might have to look for a different chip.
Leo..
ttlsksky:
Wawa,
I want to use the video source (24x24 pixels) to operate 576 electromagnets.
Also, I want to adjust the magnetic force by PWM.
Link doesn't work.
Looks like one magnet per pixel. If they'll all be independent like that, split your "image" into banks of an amount of magnets that one processor (and it's associated circuitry) is able to handle, and feed your video signal into all of them. Each processor would be programmed to pay attention only to the part of the video signal that corresponded to its bank of magnets, and ignore the rest of the signal.
You can get PWM driver ships though. The one I remember is the TLC5940 which is a 16-channel 12-bit constant current sink LED driver that is porobably way overpowered for your needs, but something similar to it can be used to offload the PWM generation from the processor. With the right set of driver chips, you may be able to daisy chain them to only need one processor.
You also aren't limited to just logic level MOSFETs. It's very hard to find FETs with good 3V performance that aren't in surface mount packages, but you're not out of luck. Higher gate voltages can be controlled with a gate driver chip that sits between the FET gate and your control signal. With the appropriate gate driver, even a MOSFET with a 10V threshold can be controlled by a 3V logic signal.
In fact, since you're going to be PWMing these FETs, I would argue that gate drivers are necessary. Most digital IO pins won't have the current capability to stand up to the abuse of loading and unloading charge from a MOSFET gate frequently, and adding a resistor just turns it into a low-pass filter, which increases the FET's switching time and makes it dissipate more power. You don't want that.
Jiggy-Ninja,
Thanks to you, I first knew about 'gate driver'. But I'm afraid it costs too much.
The FQP50N06 has a voltage drop at 5V, but it has a very low Rds0n resistance.
So I want to connect 4 ~ 5 arduino uno to PC at the same time (USB) and do multi serial communication.
This prototype used a small electromagnet (120ea / 5v / 0.5a) so the TR alone was enough.
Just use a HCT type buffer/level shifter between the Arduino and the shift registers, and run them on 5volt.
You need a buffer anyway for that many 595 chips.
Leo..
Dear god, if you're assembling 576 of them, I would totally get SOT-23 MOSFETs and make custom PCBs. SOT-23 isn't bad if you have access to a reflow oven (the modded toaster ovens are fine) - I would take reflowing 576 boards with a SOT-23 on it over having to painstakingly hand solder 1628 through-hole leads any day. Just shoot me.
Also, as others have noted, you can't have an arbitrary number of shift registers on one micro, because that clock line is connected to every one of them. You have much higher fanout with CMOS than in the old days (when people talked about fanout, because it was a problem), but not that high...
I think it's not the fanout sink current of the old TTL chips had that's the problem.
It's more the total wire and port capacitance.
71 74HC595 chips is a large footprint.
SMD fets seems the solution, but don't forget the 567 * 1.3Amp current.
When did you last send 737Amp across a circuit board.
Leo..
I would have used a higher supply/solenoid voltage.
48volt is 1/4 of the solenoid current.
That voltage/current could have been done with a TPIC6*595.
74HC595 and eight mosfets in one chip.
No external mosfets needed.
Leo..
Wawa:
12volt/737Amp
8.85kW
I hope you know what you're doing.
Yeah, that might make your cellphone wig out, or your pacemaker to go into tripletime
Fortunately, 576 electromagnets do not work at the same time.
I will program to avoid that. (16 ~ 20)
In addition, the SMPS 12v 29a will be divided into twelve.
The resistance value of one electromagnet is 10 ohms.
Using 48volt, is not the current value 4.8a?
I want to reduce heat.
Today, I bought some sot-23 mosfet. Wow ... I am afraid to soldering.
I offer board with 32 N-channel MOSFETs per board that could sink that current per MOSFET.
You'd need 18 boards, would take a little while to build them all up.
http://www.crossroadsfencing.com/BobuinoRev17/
Why was 3.3V being used as the driver? I'd get that up to 5V to drive 18 boards, and maybe add some buffering to just drive them in groups of 4 so the SCK and MOSI lines are only seeing 16 loads plus interconnecting wire each. Maybe even just 3 boards.
Like a 74ACT244N, drive all 8 inputs with SCK and each output to 2 or 3 boards in series.
Buffer MOSI/Serial Data out between groups of boards as well.
CrossRoads,
I want to control each PWM.
To control 576 PWMs, a clock speed of 84Mhz is required.
Arduino due is 84Mhz but operates at 3.3v.
My electromagnet operates at 10 ohms, 576, 12v, and 1.2a. (Each PWM)
Can it be used as a board you have created?
If so, I'd like to buy it now ... I have a headache ...
I've found now is a shift register called 'HEF4094BP'
It is supply voltage is 4.5 V ~ 15.5 V.
Can mosfet with Vgs max 10v be available?