Sweet. Can't wait..............R

Hoooray... first major milestone done; here's the quick hack data output

green: pulse width
red: current
blue: voltage
purple: time ticks

More data - this time, driven from the new PLD-BACALL boost converter for DLGs.

Notice how the servo current goes from peak-to-zero many many times, only when I'm REALLY stalling it about 2/3rds of the way in the log does it stay high. At the end I swap the servo for a resistive load.

Yet more updates - had to fix the ADC reference and buffer it ... changed the source code slightly.

If you're wondering, the sampling occurs just as the servo pulse comes in. I can only just cram in enough samples (about 50 for volts, 50 for current) each time around before the next frame starts, so in effect it's an "entire frame" sampling.

A very real problem has surfaced, the lack of logging time - I think I'm going to need a 1MBit chip and that'll last about 20 minutes So I need to now make things more complex and conjure a compression algorithm.

I think I'm going to change the algorithm slightly and make the unit actually have an N-sample window with peak detection, that way it'll smooth out the rapid jitters on the current graph while still retaining the peak current but also meaning that the datastream can become more compressible.

Paul.

GOOD NEWS... no need for me to do any fancy math... I just found 16MBit replacements, that'll solve everything and give us hours of logging. All fixed

Nice going Paul. This is exciting stuff. Can't wait to be looking at my own data.

Agreed, it's going to be useful/exciting - I know I'm already thinking of using one or two of these for my own workshop.

The 16MBit chips won't be here for a few days though unfortunately, and when they arrive I have to change the firmware a little bit (different addressing scheme due to it being 16Mbits).

Paul.

The memory chips seemed to be having delay issues in delivery, so I've ordered some 4Mbit ones instead now - this will give ~80 minutes of data, so that's all still good.

Memory chips arrived today - however I was busy filling existing orders so I wasn't able to get to soldering them on or adjusting the firmware to use them.

Originally Posted by PLD

The memory chips seemed to be having delay issues in delivery, so I've ordered some 4Mbit ones instead now - this will give ~80 minutes of data, so that's all still good.

80 minutes works for me (as long as that's not a lifetime limit :-). I'll just get 80 minutes of data, then do it over and over. Bring it!

I put in the new memory chip... hoping that everything will be ready to run tomorrow. Amazing how long these "simple" things can sometimes take to finish off and get out the door.

Originally Posted by PLD

I put in the new memory chip... hoping that everything will be ready to run tomorrow. Amazing how long these "simple" things can sometimes take to finish off and get out the door.

Hi Paul -- Howgoezit with the "ready to run tomorrow" ?

BoB R.

I ran into some very "odd" behaviours. The whole unit started oscillating with reboots, most confounding / frustrating. I added more decoupling capacitance to the regulator stage which seems to have assisted in eliminating the fault but it's still not correct behaviour, hence I'm trying to nail down the true cause. I have some replacement boards already en-route, which I hope will clean the issue up.

Paul