Connecting load to circuit resets MAX809

Question

I have a MAX809T which partially works fine, outputting high when voltage is above the threshold after a small 1 second time delay, and outputs low when the voltage is below the threshold. It also has a hysteresis. However, when I connect a load, even a small 70mA microcontroller, the MAX809T actively pulls its output low, waits the 1 second time delay, then pulls it back up.

The battery voltage is 3.99V, and it doesn't change (to 0.01V resolution of my multimeter), and the MAX809T threshold is 3.08V.

I don't believe a small microcontroller can draw a large enough start-up current to drop a large LiPo cell down 1V, but even so I tried adding an electrolytic 10uF capacitor but it doesn't change this unexpected behaviour.

I even tried adding a pull-up resistor to pull the output high, but the MAX809T is actively pulling it low so this doesn't work.

I don't have any external circuitry connected to the MAX809T, just VCC and GND connected, with a multimeter measuring between RESET and GND.

Multimeter shows the average of a number of measurements so you may not be able to catch any sudden and temporary drops. Try with a scope instead. Also you may want to monitor the input of your supervisor IC to see if it's noisy. — Rohat Kılıç, Jun 06 '23 at 10:49
I don't have a scope, but I will assume this is the case, as per @Spehro Pefhany's answer — UnknownDagger, Jun 06 '23 at 11:18

Spehro Pefhany · Accepted Answer · 2023-06-06T10:59:49.880

3

The MAX809 should have a small ceramic (not electrolytic) bypass capacitor (perhaps 0.1 to 1 uF) mounted very close to its supply pins.

Most likely your “70mA” microcontroller has bypass capacitors so when you close the switch there will be a negative-going transient on the power supply of tens or hundreds of microseconds which will trigger the MAX809 reset pulse (indeed, as it should). This is exactly what it is designed to do, since such a transient could cause an upset to digital circuitry.

If you want it to do something else, perhaps you could isolate the supply of the AD809 from the battery with a series resistor and provide it with a substantial bypass capacitor at the chip. Pull-up to the battery side of the series resistor. That will prevent it from detecting the disruption caused by the switch closing. An oscilloscope would be helpful in getting the numbers correct, or detailed knowledge of what is being switched. Without that it’s just guesswork, but the supply current of the AD809 is quite low, making this approach possibly feasible without too much error in the detection voltage.

edited Jun 06 '23 at 10:59

answered Jun 06 '23 at 10:50

Spehro Pefhany

376,485
21
320
842

There is no capacitor in the MAX809's datasheet so I wasn't aware I needed one. Thank you – UnknownDagger Jun 06 '23 at 11:19
Which is best, 0.1 or 1 uF, or even 10uF? – UnknownDagger Jun 06 '23 at 11:21
0.1uF is plenty for bypass purposes, however it's not likely going to solve your problem. 1k series with the power supply and 470uF in parallel with the 0.1uF may do so. – Spehro Pefhany Jun 06 '23 at 11:30
you mean a 470uF electrolytic? Maybe I ought to use a different component, this is far too work-aroundy! – UnknownDagger Jun 06 '23 at 12:20
Yes, it's designed to very reliably reset MCUs with any power disturbance and it sounds like you just want a reference and a comparator. – Spehro Pefhany Jun 06 '23 at 12:30
exactly, I was looking for one with low power consumption and a threshold around 3.08V, do you have any recommendations? – UnknownDagger Jun 06 '23 at 13:40
I would probably use something like LMV431, a couple resistors, and an MCP601 op-amp as a comparator. – Spehro Pefhany Jun 06 '23 at 16:59

Connecting load to circuit resets MAX809

1 Answers1