azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing | https://github.com/azonenberg/scopehal-apps | Logs: https://freenode.irclog.whitequark.org/scopehal
juli966 has quit [Quit: Nettalk6 - www.ntalk.de]
electronic_eel has quit [Ping timeout: 240 seconds]
electronic_eel has joined #scopehal
Degi_ has joined #scopehal
<d1b2> <david.lenfesty> azonenberg: do you have a standard TVS diode you use for digital signals?
<d1b2> <david.lenfesty> this checklist is full of things I usually leave off my designs because I'm far too lazy :p
Degi has quit [Ping timeout: 240 seconds]
Degi_ is now known as Degi
<azonenberg> david.lenfesty: let me check, i've used a couple
<azonenberg> SP4010-02HTG?
<d1b2> <david.lenfesty> LGTM
<d1b2> <david.lenfesty> I really need to come up with a good system for inventory and just listing common parts I use
<_whitenotifier> [starshipraider] davidlenfesty synchronize pull request #4: Active Probe Host Prototype - https://git.io/JLHJu
<azonenberg> david.lenfesty: at this point are you done with schematic then?
<azonenberg> and yes, the checklist covers lots of things people often leave out. But the goal is to create designs that are right the first time, as much as possible
<d1b2> <david.lenfesty> Yeah I'm done unless I did something dumb / I'll probably change some FPGA pins to make routing easier once I get to that.
<d1b2> <david.lenfesty> azonenberg
<azonenberg> Great
<azonenberg> TiltMeSenpai: What's the status of your probe-side power stuff?
<d1b2> <TiltMeSenpai> I haven't poked the design much, I'm still trying to figure out what's the best way to handle input protection for a rogue usb c charger
<d1b2> <TiltMeSenpai> I can do an initial revision with minimal input protection though
<azonenberg> That's probably good for starters
<azonenberg> From my perspective we only need enough of a probe to verify the host
<azonenberg> The host needs to be finalized before we build BLONDEL
<azonenberg> the probe side can evolve, as long as it doesn't break hardware compatibility with the host
<azonenberg> Once we build BLONDEL the host side will be at least somewhat frozen
<d1b2> <TiltMeSenpai> ok, the 2 hold ups were deciding on input protection and figuring out how to safely detect -7v (I haven't thought of that as much)
<d1b2> <TiltMeSenpai> given that a few people have claimed to have usb c cables without SS lines, I think the probe needs to report a "+/-7v missing" error state
<d1b2> <david.lenfesty> well do we need to detect -7V?
<d1b2> <david.lenfesty> Are there any cables with SSTX but not SSRX?
<d1b2> <TiltMeSenpai> oh, that's kinda a good point. We could certainly cheat and only detect +7v with a simple resistor divider, but that feels like cheating
<d1b2> <david.lenfesty> I wouldn't call it cheating. I would call it "cost optimizing" 😛
<d1b2> <david.lenfesty> But is there any reasonable scenario where we have +7V and NOT have -7V?
<azonenberg> PSU failure on the host
<azonenberg> or damaged cable/connector
<azonenberg> Both would be good to detect
<azonenberg> I think you could easily detect -7V with a voltage divider
<d1b2> <TiltMeSenpai> can the stm32 handle negative voltages?
<azonenberg> It should be able to handle less than one diode drop below ground
<azonenberg> Basically the idea is, set up a center tapped divide with a resistor to ground
<azonenberg> such that +7 and -7 gives 0V at the midpoint
<azonenberg> +7 only gives a positive voltage
<azonenberg> -7 gives a negative voltage
<d1b2> <TiltMeSenpai> oh see if v+ - v- is 0
<d1b2> <TiltMeSenpai> I see
<azonenberg> Then have a separate sensor for the +7 rail to distinguish the no-power from the both-power state
<azonenberg> You could do it with two taps on the same resistor ladder and two adc inputs on the stm32
<azonenberg> Check the stm32 datasheet. as a general rule, <0.5V beyond a power rail is normally safe
<azonenberg> because the protection diode starts conducting at around 0.5 to 0.7V beyond the rail
<azonenberg> Actually i think if you are clever with the resistor values you can do it with one tap
<azonenberg> choose unequal tap dividers
<d1b2> <TiltMeSenpai> I don't know if that will end up handling equal v+/v- that aren't 7v but do we care as much about that?
<azonenberg> Hmm, good point
<d1b2> <TiltMeSenpai> actually none of my solutions handle that either so
<azonenberg> I do want monitoring especially on early rev probes
<azonenberg> ideally i want to see voltage drop on the rails
<d1b2> <TiltMeSenpai> do we need an external 2 channel differential adc then?
<azonenberg> Well the bigger question is how much can we rely on it if we don't know whether the external power is present?
<d1b2> <TiltMeSenpai> or can this be done with an opamp, I'm not as familiar with all the operating modes of opamps
<azonenberg> We might still end up needing two separate solutions
<azonenberg> One using the MCU ADC to detect on/off and measure the positive rail
<azonenberg> then a separate optional i2c split rail adc that we can use to measure the negative rail voltage
<azonenberg> and DNP if we decide it's not needed
<d1b2> <TiltMeSenpai> is it ok if I just break out the +/-7v to a 2.54mm header while we think about this
<d1b2> <TiltMeSenpai> I'll give a couple adc capable pins on the same header, we can throw the detection circuit on a breakout board for prototyping that way
<azonenberg> For the prototype i suggest test points (2.54mm header is OK i guess, but Keystone 5016 would be my preference) for each
<tnt> azonenberg: by anychance do you remember the part number for the connector on the solder-in probe on flex you did ages ago ?
<azonenberg> tnt: the MMCX?
<azonenberg> It was a Linx right angle SMT MMCX, i think they only make one
<azonenberg> digikey search should find it
<tnt> Ok, thanks !
<d1b2> <TiltMeSenpai> there should be space for both footprints but I have 2.54mm headers on hand lol
<azonenberg> Lol fair enough. I just like the keystone connector because it's easy to grab onto and has more clearance around it
<azonenberg> especially for higher voltages you dont want to risk shorting them together
<_whitenotifier> [starshipraider] TiltMeSenpai forked the repository - https://git.io/JL559
<azonenberg> tnt: incidentally that probe was v0.1 of the AKL-PT2
<azonenberg> It just wasn't called that at the time
<tnt> azonenberg: yeah I figured it was a pre-cursord. So far I've literally been using a 450 resistor soldered at the end of a coax, so I figured I should start using the ones you sent :p
<d1b2> <TiltMeSenpai> any preference with rgb status lights or just 0805 red LED's on the probe prototype?
<tnt> I'm sure the AKL-PT2 is now much better but ... I also only have a 500 MHz scope so I'm not sure I'd be able to see the improvements anyway.
<d1b2> <TiltMeSenpai> I have a ton of 0505 RGB LED's and they're fun
<azonenberg> The old one is probably fine to a few hundred MHz but the connector launch is not well matched
<azonenberg> You might still see a benefit with the PT2 because it's got a flatter input impedance. So it loads the DUT less
Bird|ghosted has joined #scopehal
Bird|otherbox has quit [Read error: Connection reset by peer]