Or no resistor, no measurement. The uP could drive the relays ON at 100% duty cycle initially, but then drop down to something like 70% after a reasonable pull-in delay.
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
I was thinking of a ULN2003 hex darlington array, with built in flyback diodes, would work: One hex section per relay. One spare section. No extra parts, no regulator, no added diodes. Slightly easier wiring if the relays are scattered all over the PCB. Approx 200-300 milliwatt device dissipation when relays are energized.
At 3.0V from the GPIO pin, input current is about 500ua which 3.3V CMOS should handle. If it's a Rasberry Pi, the default GPIO output pull up or down current is 8ma, which is more than enough.
I don't see any advantage to this scheme, except that it makes troubleshooting a bit easier since the relays can be energized individually.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
I don't have an example. When working with automotive type relays, I assume that they will work within 10% of nominal voltage (13.6 VDC) without difficulty, 15% to the limits of the maximum ratings, and 20% if you're lucky and can control the temperature. Obviously, this varies with the situation, but has been a good general rule for first approximations.
In the marine radio biz, marketeering decided that the radios needed to operate on the battery running the trolling motor on small motorboat used for fishing. Normally, it's a half dead automobile battery quickly charged before going fishing. Some crude measurements showed that such a derangement was likely to bring the battery voltage down to as low as 10.0 VDC[1]. Since the radios were for emergencies, it was assumed that if it was that low, the boat was having a genuine emergency. Ignoring the improbability of the scenario, marketing was deemed always right by management, and must be obeyed. Note that this exercise was to be performed on the entire line of products (about 10 different model radios at the time).
So, I took a radio that was designed for 11.5 to 15.5VDC (+/- 15%) and tried to squeeze another 10% out of the lower end of the voltage range. This was initially quite a trick because the internal regulators were all set to 9.1 VDC. After the regulators were convinced not to oscillate (and later replaced with LDO regulators), the audio amplifier stopped motorboating, making the Class C stages stay in Class C operation, and some bias points moved, I found that the T/R relay was having a problem. Because of the low turn on voltage to this relay, it was taking longer than normal to energize. During this delay, the transmitter would be putting out full power into an open circuit. That would cause the VSWR protection circuit to kick in, which would eventually cause the relay to play buzzer. The resultant arcing across the contacts usually fried the relay or produced a high resistance connection.
It was easy enough to replace the 12 VDC relay coil with an 8 VDC equivalent, except that the coil would fuse open during burn-in at
15.5 VDC. So, I added a current source in series with the 8 VDC relay, and wasted some power. That worked.
I don't recall exactly what voltage the T/R relay was running on, but my guess is that it was fairly close to 11.0 VDC with 13.6 VDC applied. Between the battery and the relay were numerous connectors, power cables, two fuses, on-off switch, internal wiring, transistor PTT switch, etc. At 50-100 mv drop per connection, these voltage drops add up. Also, at whatever current the transmitter was pulling (about 6A in low power), the voltage drop only gets worse.
When I first saw the 12v to 18v power spec on the faulty schematic, I tried to guess what was powering this contrivance. My best guess is a wall wart transformer, diode bridge, and minimal filter capacitor. No regulation. That's not the best power source available and is likely to cause problems unless the characteristics of the wall wart are known. I would think that regulating the wall wart voltage might be better than regulating just the relay voltage.
[1] Despite my bad attitude towards making the radio run at 10.0 VDC, it later became an important selling point and was maintained as a key design spec because none of the competing products even came close.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
One should also consider the expected ON/OFF ratio. If it is 50%/50% or unknown, it does not matter, but if it is 1%/99% or 99%/1% the driver circuit should consume current only at the shorter period and select topology (PNP/NPN CE/CC) accordingly.
This topology also determines, if it can be driven by ordinary 3.3/5 V TTL or if some special 15/30 V tolerant open collector drivers are needed.
In that case, use a PNP transistor at 12 V, the base goes through a resistor to a 15 V tolerant open collector TTL output or to a NPN transistor driven by a 3.3 V controller pin through a resistor.
One should remember that the relay turn on current is greater than the hold on current.
Radio amateurs sometimes use "24 V" relays with 12 V car battery by charging capacitors to 12 V at off state and when switched on, the capacitor is in series with the 12 V feed, thus producing 24 V, but of course, after a while, the voltage drops to 12 V, which is sufficient hold for most 24 V relays.
In this case, I would not bother to regulate the 12-18 V input at switch on (let the transistor be hard on), but after a while drop the relay voltage to something like 6-9 V, which should give a sufficient hold current. This will reduce the relay dissipation significantly. Some of the dissipation moves to Q2 at steady state, but as the total current drops, the total dissipation also drops.
PWMing 200 mA though big inductances seems to me as asking for troubles, at least in EMC compliance testing, if not earlier.
If the processor only has digital outputs, then PWMin might be OK, but I would still keep Q2 as a linear regulator, dissipating a lot of power and low pass filtering the base drive to remove most of the PWM waveform.
Best for each relay to have a series resistor, and a diode in series with resistor across the coil. The diode-resistor combo is to snub flyback; the resistor decreases snub current and thus allows faster relay release; prolly 30-200 ohms. That coil-resistor/diode network should have that series resistor mentioned, and should drop 5-10% of total (5V) voltage applied; these tend to balance coil drives and isolate spikes - which decreases what i call phase adding (more like EMI adding like crazy).
As a rule you can assume the relay pull-in voltage is just equal to the nominal coil rating at the maximum of the spec'd operating temperature range. This will be due to increases in magnetic circuit reluctance as well as winding resistance.
The PWM frequency can be kept low, a few KHz maybe. Any EMI will be at the 200th harmonic or something insane like that, so the spectral energy will be minute. You wouldn't expect any EMI trouble from a 300 mA buck switcher running at 5 KHz.
Schematic?
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
A depletion fet is normally ON and acts like a constant-current device. It acts like a resistor to pull up the base, but the current doesn't ever go high as it would if you used a resistor.
An LND150 would regulate to maybe 1.5 mA, so you'd need a transistor with a beta of 200 min. There are other depl parts with higher Idss.
Not all of my circuits are always guaranteed to be non-silly, especially when I do seven in a couple of minutes for free. The next step, for any circuit, is to apply math.
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.