ESR meter upgrades

[stj]

Re: ESR meter upgrades

most people use stripboard on wires

[eccerr0r]

Re: ESR meter upgrades

Finally got my avrtransistortester... took enough time, so much for china shipments.

It came with a crap solderjob, had to reflow it before the display (rgb color) would work.

Pretty interesting. device. Crappy slow UI... wonder if other versions are as slow...

Now...3d printed case for the device, hah...

[stj]

Re: ESR meter upgrades

color display slows the UI because it takes 3 times the data to be sent over the spi bus for each refresh.
if you change the crystal to 20MHz and re-flash the firmware it runs much better.

[eccerr0r]

Re: ESR meter upgrades

TBH I've seen HD44780 updates really slow too, and that's parallel even - except written/designed poorly. I just wonder if the same is with AVRtester or not, lazy programmers or not...

My PSU project I was working on with the HD44780 display, designed/written "properly" I was able to get it to update autorepeat keypad buttons on the screen without it skipping, and still change the numbers faster than the numbers can be read. Then I see other implementations that can only update once per second... Boo!!!

Now I wonder how long these things last with a PP3 battery or is it not worth it, and have to use a Li-ion pack...

[stj]

Re: ESR meter upgrades

the pp3 lasts forever - it can run one down to about 6v and still function fine.

HD4470 is an interesting case,
it has a "busy" pin/flag.
if you watch it between transfers it runs fast, but a lot of designs save a pin by just using fixed delays between writes.
also it can run in 4bit or 8bit mode

[redwire]

Re: ESR meter upgrades

HD44780 runs really slow, I think its internal clock is a few hundred kHz. The 1,000nsec (data) cycle time gives it away. I used the Busy Flag once and it was so slow and F/W could hang if the display did not respond, they do crash due to ESD.

I 'd like to see the ESR Meter ported over to something ARM-core like SAMD21G18. 256KB FLASH, 32KB RAM, 48MHz.

[eccerr0r]

Re: ESR meter upgrades

Well, thought the HD44780 was quite fast or at least "more than fast enough" when the code/interface is written right... and if lazy it's a slug. Hoping that this SPI color LCD interface, even if it's only 16-bit, is not affected the same way.

[stj]

Re: ESR meter upgrades

well the hitachi displays are character displays - so even the biggest (40 x 4) only needs a 160byte packet for a complete refresh.

the spi displays are bitmapped 160 x 128 or bigger.
so 3bytes per pixel = 6K per refresh.
to make it worse, most testers dont use the hardware spi or i²c pins - they bitbang the data.

as for porting to an ARM cpu, the problem is the cpu needs to be running at 5v.
unless you want to add seperate ADC / DAC chips.

[eccerr0r]

Re: ESR meter upgrades

I just think there's no excuse to send 8 characters per second to the HD44780, just to save a pin/lazy code!

[stj]

Re: ESR meter upgrades

I just think there's no excuse to send 8 characters per second to the HD44780, just to save a pin/lazy code!

that would be classed as a piss-take!!

[eccerr0r]

Re: ESR meter upgrades

hmm... playing with transistortester... found a few parts that it doesn't seem to work or maybe my parts are broken... was there any firmware history that perhaps a newer firmware would detect:

- I have one TO220 TRIAC that it thinks is dead, though I thought this TRIAC was NIB. Hmm. I have another, very low power TRIAC that it does detect.
- A unijunction transistor, it detects as two diodes.
- Darlington transistors like TIP125, beta is very low. Vbe is also lower than expected.
- A lot of other TO220 BJT return very low betas (sort of expected at high current however, but how much current is it testing?). TO92 and TO39 transistors, beta is much higher.

I had this one TO5 transistor that was really weird as it was open all around. I concluded without the transistortester that it was a E-MOSFET. So I finally gave it a go and it agreed with me that it indeed was an E-MOSFET. It seems to correctly identify N-JFET that I had as well.

Found a few bipolar capacitors that probably needs to be binned at 11Ω ESR...

Generally an interesting tool...

[stj]

Re: ESR meter upgrades

triacs dont work - they need too much current to turn on

[eccerr0r]

Re: ESR meter upgrades

Interesting, yeah kind of figured it couldn't supply much current. Still wonder why darlingtons show up real low in gain however.

It could detect thyristors, but had problems with a few to220 thyristor as well.

It detected one of my BT136's as a BJT with very high hFE, and another as (unidirectional) thyristor... Can't trust this the tester for high current devices it seems, have to go back to manual hand testing.

[stj]

Re: ESR meter upgrades

i have tried a bunch of triacs and sometimes they are id'd as scr's other times as transistors.
i suspect it's down to how the triac reacts at low current.

[eccerr0r]

Re: ESR meter upgrades

Have to admit this is kind of disappointing that it can't detect high power thyristors and it's even frequently wrong...

Oh well.

[stj]

Re: ESR meter upgrades

6mA is the test current.

[eccerr0r]

Re: ESR meter upgrades

At least it does OK measuring ESR

Unfortunately it does seem to get tricked by leaky capacitors. Tried making an artificial leak and the capacitance went up...

[stj]

Re: ESR meter upgrades

but the full software shows the leakage too

[eccerr0r]

Re: ESR meter upgrades

yeah, at least it displays loss, but it's still wrong Seems like it will still try to generate a capacitance reading, as long as it doesn't leak way too fast...

[redwire]

Re: ESR meter upgrades

You need a high current pulse to test ESR with bigger capacitors. If a good portion of that pulse is lost in the leakage current, then the ESR reading is wrong but capacitance should be wrong as well (if it's using single-slope technique).
The Anatek Blue ESR meter has three constant-current test levels - 0.5mA, 5mA and 50mA which works well. It's only an 8usec wide pulse, and check is on the ramp up and down.

ESR meter - it might be the firmware's timing for when to apply the pulse and then sample with the A/D? Maybe the gap is not extended with the faster MCU crystals? Not sure how it measures ESR.
chinese electrolytics act more like leaky batteries, tough to figure out if the part works lol.
I think this tester needs to get out of cheap town and somebody blow the budget and add a transistor lol, so it can pump some decent test current out - into a capacitor or diode or whatever. I thought it's just 5V/680R or under 7mA but the MCU's output drop is also in the way.