#robotics Logs
Jul 22 2018
#robotics Calendar
12:00 AM rue_shop4: mrdata,
12:04 AM rue_shop4: I have one for ya to play with simulation
12:04 AM rue_shop4: take a cd4011
12:04 AM rue_shop4: tie one input of each gate to the chips vcc
12:04 AM mrdata: X[C] = 1 / 2 pi f C = 3 k ohm
12:08 AM mrdata: 1 /( 2 pi f C ) that is
12:08 AM rue_shop4: tie output 1 to input 2, output 2 to input 3, output 3 to input 1 VIA a 2.2M resistor, and output 3 to input 4
12:08 AM rue_shop4: varry the vcc from 2V to 12V
12:08 AM rue_shop4: I get a result freq from 500Khz to 26Mhz
12:08 AM rue_shop4: no capacitors
12:08 AM rue_shop4: only adjusting vcc
12:08 AM rue_shop4: (signal off output 4)
12:08 AM rue_shop4: if you make a ring osc with the 4069, I mere inverter, its NOT as fast... I dont get it
12:10 AM rue_shop4: strapped a 470 ohm resistor to the top of it...
12:11 AM rue_shop4: applying 5V causes the freq to go down
12:12 AM rue_shop4: yea, wow
12:12 AM rue_shop4: verry temp sensitive
12:15 AM rue_shop4: I wonder how casting it in wax would work for stability
12:48 AM mrdata: remind me what the final application of this is
12:58 AM rue_shop4: OMG 15Khz to 50Mhz
12:59 AM rue_shop4: thats insane
12:59 AM rue_shop4: bench signal gen
12:59 AM rue_shop4: the only problem is that at 5V levels, the ttl rings like CRAZY
01:01 AM mrdata: this is fixable
01:01 AM rue_shop4: maybe I should take some pics...
01:02 AM mrdata: what frequency is the ringing
01:03 AM rue_shop4: well, its over 45Mhz
01:04 AM rue_shop4: 1 cycle in ~ 0.02us
01:04 AM rue_shop4: so 50Mhz
01:05 AM rue_shop4: .... maybe I should run the buffer at 3.3V?
01:05 AM mrdata: ok, so an RC filter with appropriate cutoff is advised. tau > 0.02 us
01:05 AM rue_shop4: instead of 5
01:07 AM mrdata: 1 ohm and 30 nF, say
01:07 AM mrdata: or 10 ohm and 3 nF
01:07 AM rue_shop4: latter sounds more reasonable
01:08 AM rue_shop4: but can a 74hc04 switch at 10Mhz without ringing?
01:09 AM mrdata: test and adjust. if you see ringing, slap an RC filter on it
01:13 AM rue_shop4: ARG, the freq counter was just parasitically picking up the signal
01:13 AM rue_shop4: I didn't have it attached to the circuit
01:16 AM rue_shop4: ok, still insane
01:17 AM rue_shop4: it'll go from 10Mhz to 150khz
01:19 AM rue_shop4: stability establishes about .. oh sorry 3.8Khz
01:20 AM rue_shop4: so 3.8khz to , well about 11Mhz
01:20 AM rue_shop4: I wonder what the freq counter was picking up the 50Mhz from
01:21 AM mrdata: outer space
01:21 AM mrdata: maybe something closer
01:21 AM rue_shop4: at a thres of 0.02V?
01:22 AM rue_shop4: and in sync with me dialing this osc?
01:22 AM rue_shop4: there are no caps in it, its just 3 NAND gates and a 2.2M resistor
01:52 AM rue_shop4: apparently a 220pF load cap cleans it right up
01:53 AM rue_shop4: if my plan is to drive it thru a 50R probe to the target circuit, I wonder how thats going to play out
02:11 AM rue_shop4: wtf, I'v had these scope probes for years, and didn't know they pull apart to reveal a coax connector
02:11 AM rue_shop4: damn
02:15 AM rue_shop4: huh, yea, transporting 10Mhz is gonna be trickey
02:27 AM Jak_o_Shadows: can you just do it with nicer cables?
02:46 AM rue_shop4: whats nicer than a 50 ohm scope probe?
02:46 AM Jak_o_Shadows: i dunno
02:46 AM rue_shop4: rated for 60Mhz
02:47 AM Jak_o_Shadows: but like, I know that you can use normal coax for high frequency signals
02:47 AM rue_shop4: its too big
02:47 AM Jak_o_Shadows: ok, remove "normal"
02:55 AM rue_shop4: ok I'm gonna rewind my design a bit
02:55 AM rue_shop4: I think I need to clean up the signal off the osc more before I try to do anything with it
02:55 AM rue_shop4: I need an amp with a flat response from 1Mhz to 10Mhz
02:56 AM rue_shop4: well, no, higher, its square
02:56 AM rue_shop4: or is it
02:56 AM rue_shop4: the 2222 is rated to 250Mhz
02:56 AM Jak_o_Shadows: fat in which way? phase?
02:56 AM Jak_o_Shadows: magnitude?
02:57 AM rue_shop4: well yea, I need to get the signal level up more
02:57 AM rue_shop4: maybe I can throw some agc at it
02:57 AM rue_shop4: its only 1am and I'm getting all woozy
03:06 AM Jak_o_Shadows: What sort of gain do you want?
03:22 AM rue_shop4: I think I got it, but hmm
03:22 AM rue_shop4: how do I know if the slew I'm seeing is the circuit or my scope
03:22 AM rue_shop4: 25Mhz scope, 10Mhz signal
03:23 AM rue_shop4: showing my a slew of 2V/0.02us
03:24 AM rue_shop4: but I had to load the last chip with a 220pF to ground to quench the ring
03:24 AM rue_shop4: hmmm
03:31 AM rue_shop4: dunno if a 74hc04 can even switch clean at 10Mhz
03:32 AM rue_shop4: datasheet says 6ns rise/fall
03:32 AM rue_shop4: I'm at 20ns
03:34 AM rue_shop4: the HCT244 is supposed to be blindingly fast
03:49 AM rue_shop4: odd
03:50 AM rue_shop4: dispite better slew and propigation numbers, a ring osc made with the 74HC240 does not compare to the one made with the cd4011
03:50 AM rue_shop4: the 240, with only 1 gate, can only get to 1.2Mhz
03:57 AM rue_shop4: the 74HC00 can almost get to 30Mhz on a breadboard
03:57 AM rue_shop4: with a low of 56Khz
04:15 AM rue_shop4: the signal from the osc is sloppy, what I'm hoping to do, its throw so much gain at it, that it distorts itself into a square wave
04:15 AM rue_shop4: one with nice edges
04:16 AM rue_shop4: I thiink the input capacitance of the pn2222 is killing my efforts
04:22 AM Jak_o_Shadows: I really need to get a little drill
04:22 AM Jak_o_Shadows: doing it by hand is very annoying
04:23 AM Jak_o_Shadows: or I should have put it into the 3d model when I printed it
04:39 AM rue_shop4: heh
04:39 AM rue_shop4: mcp6002
04:39 AM rue_shop4: 1Mhz
04:42 AM rue_shop4: ah 6482
04:43 AM rue_shop4: 1.5Mhz, grrr
04:45 AM rue_shop4: aha, NE592
04:45 AM rue_shop4: oh idea
04:48 AM rue_shop4: the way I did this gives me differential output from the osc
04:49 AM rue_shop4: oh hey, is an RS422 reciever rated for 10Mhz?
04:50 AM rue_shop4: bah 250kbs
04:50 AM rue_shop4: hmm
04:54 AM rue_shop4: 75175 is a no
04:54 AM rue_shop4: aha 3468
04:54 AM rue_shop4: ... max3468 up to 40mhz
05:04 AM Jak_o_Shadows: http://pasteall.org/pic/show.php?id=fe7c0860c03276e4abf4808629526b96
05:04 AM Jak_o_Shadows: base platform (the white) is a bi flimsy, would need reinforcing
05:04 AM Jak_o_Shadows: but it works for now
05:05 AM Jak_o_Shadows: backlash seems to be coming more from the servos than else
05:09 AM * rue_shop4 dances
05:09 AM rue_shop4: hahahaha
05:09 AM rue_shop4: got it
05:09 AM * Tom_L checks his watch
05:09 AM rue_shop4: ok, I made the oscillator centre on the 5v supply, then I use a RS422 reciever to translate the differential signal to 5v
05:09 AM Jak_o_Shadows: I have maybe 30 degrees max change in angle? That's enough for this use.
05:09 AM rue_shop4: 3am
05:10 AM rue_shop4: 30 degrees, not sure
05:10 AM rue_shop4: havn't sat down with the math on those
05:21 AM Jak_o_Shadows: The maths is complicated
05:21 AM Jak_o_Shadows: Well, that maths is anyway
05:29 AM rue_shop4: it looked like fun, I was going to sit down with it at some point, might still sometime
05:29 AM rue_shop4: just a union of radii really
05:29 AM rue_shop4: ok
05:30 AM rue_shop4: so the max3468 works nicely, but the output of it rings too
05:30 AM rue_shop4: and I can stabalize my 74hc04 based amp better than the max chip
05:31 AM rue_shop4: what I need now is a freq to voltage converter thats linear from 1Mhz to 10Mhz
05:31 AM rue_shop4: so it cant be a filter
05:31 AM rue_shop4: unless its a balanced bridge?
05:31 AM rue_shop4: er, no dont think so
05:34 AM rue_shop4: the 74hc00 is better than the 4011, it can hit the 10Mhz with a 5V rail, the 4011 takes 6V
05:34 AM rue_shop4: the hc04 works fine for cleaning the signal up, I need a driver now tho
05:34 AM rue_shop4: I'm thinking a video cathode driver
05:34 AM Jak_o_Shadows: Yeah, it basically is the union of radiai
05:41 AM rue_shop4: LM2419
05:41 AM rue_shop4: 65Mhz driver
05:41 AM rue_shop4: capable of 65Mhz at 50V p-p
06:00 AM rue_shop4: ok, that will drive 10Mhz into anything at 24V p-p, but
06:00 AM rue_shop4: its still not square
06:05 AM rue_mohr: how do I 'square up' a 10Mhz signal
06:05 AM rue_mohr: I might have to filter it it death first
06:06 AM Jak_o_Shadows: Don't you normally do that with one of those schmitt trigger things?
06:06 AM rue_mohr: cant, they ring too much at 10Mhz
06:07 AM Jak_o_Shadows: How do you get a square wave with filters? Like, filters decrese the bandwidth of your signal, a square wave has heaps
06:07 AM rue_mohr: the ring oscillator puts out a signal thats like a RC charge/discharge, so its not pretty in the first place
06:08 AM rue_mohr: then again, if I use that rs422 reiciever, its doing 40Mhz comparitoring to square up its signal
06:08 AM rue_mohr: problem being, its output rings too
06:08 AM rue_mohr: I think I need all-round better line termination
06:09 AM rue_mohr: it is reasonable to assume you cant leave a 10Mhz square wave comming off a chip pin without the right impedence load
06:09 AM rue_mohr: obviously my x10 scope probe isn't it
06:10 AM rue_mohr: on the filters, thing, you have to round off the corners of the wave
06:10 AM Jak_o_Shadows: I know it won't give you the accuracy you want... but for testing out all the other components, seeing if they're capable of carrying a 10MHz wave, you could get one of teh stm32's to toggle a pin?
06:10 AM rue_mohr: yes, and its output rings too
06:11 AM rue_mohr: I was going to make the whole thing with stm32 at one point
06:11 AM Jak_o_Shadows: Yeah, but it seems like it could be an ok point of comparison
06:11 AM Jak_o_Shadows: Just to help tell you where the mess comes from?
06:11 AM rue_mohr: if I run it at 60Mhz, and divide by 6, I get 10Mhz, but if I divide by 7, I get like 8.3Mhz, so, er, no
06:12 AM rue_mohr: toggling a pin at 10Mhz was interesting too
06:12 AM rue_mohr: those things sure do have a lot of stuff you have to enable
06:15 AM rue_mohr: it would probably fix if I had a really low-Q filter to run it thru, to sine it up, then amplifed the hell outa it till it distorted
07:31 AM rue_shop4: ok, if you follow the datasheet, the amplifier works better
12:00 PM mrdata: when in doubt, read the directions, eh
02:37 PM rue_: oops, crashed machine simulating a RLC network
02:37 PM rue_: mrdata, can you easily doa quick runoff of a parallel RLC with 20R, 200nH, and 3.8nF all in parallel
02:38 PM rue_: need to confirm if it has a ~10Mhz bandwidth
02:38 PM mrdata: ok
02:38 PM rue_: wow, I was up to like 5am
02:39 PM rue_: I think I need cmos and gates with complimentrary outputs....
02:39 PM rue_: but shower and breakfast come first
02:41 PM mrdata: rue_, theres a problem
02:42 PM mrdata: rue_, i need total series resistance
02:42 PM mrdata: including source and load
02:44 PM mrdata: and notice, the cap and the inductor have ESR; what to use for that
02:47 PM mrdata: i'm putting in 0.01 ohm for the cap, and 0.001 ohm for the inductor
02:49 PM mrdata: bandwidth is usually quoted as 3 dB band width
02:56 PM mrdata: but this filter doesnt give you 3 dB at 10 MHz band width
02:56 PM mrdata: it comes close, though, if source and load are 8 ohms each
02:56 PM rue_: hmm
02:57 PM rue_: thats probably ideal values, with a 0 ohm source... no, hmm...
02:57 PM mrdata: no
02:57 PM mrdata: with 0 ohm source, it's nuts wide
02:57 PM rue_: yea
02:57 PM rue_: 8 ohms, that would make sense, it was probably a site for an audio filter
02:57 PM rue_: hah
02:58 PM rue_: centre about 5.5Mhz?
02:58 PM mrdata: yeah
02:58 PM rue_: hmm ok
02:59 PM rue_: I spent a lot of time wrestling with proper ttl line termination for the high freq's, there isn't much out there about it
03:00 PM rue_: most consistant answer was 220R to ground, 330R to vcc, and 20pF ea up and down
03:00 PM mrdata: for TTL, if it can source/sink 40 mA, say, then it's about 125 ohm impedance
03:00 PM rue_: tho, that didn't touch most of the ringing I was fighting
03:00 PM veverak: so
03:00 PM rue_: hmm, ... whats 220pF @ 10Mhz....
03:00 PM veverak: how to measure resistance of PT100 ?
03:01 PM veverak: voltage divider or constant current?
03:01 PM rue_: veverak, water boils at 100c
03:01 PM veverak: I am interested in 100-250 ohm range
03:01 PM rue_: oh, in general, if you dont want to do the software compensation use a current source
03:01 PM rue_: otherwise reprap uses a pullup resistor and a software table
03:01 PM Tom_L: veverak we just used the 1.1v avr voltage reference
03:02 PM mrdata: rue_, X[c] at 220pF @ 10Mhz is 72.3 ohm
03:02 PM Tom_L: directly to an avr adc pin
03:02 PM rue_: drat you beat me
03:02 PM rue_: I need to make a command line for that calc
03:03 PM veverak: Tom_L: ?
03:03 PM rue_: mrdata, 220pF to ground almost perfectly took out the ringing
03:03 PM veverak: rue_: I see
03:03 PM mrdata: rue_, good
03:03 PM Tom_L: veverak for our toaster ovens we used a PT100
03:04 PM veverak: Tom_L: yeah, not sure I understand how you measure it's resistance?
03:04 PM rue_: mrdata, but I couldn't tell what was the ttl gate output and what was the test leads
03:04 PM rue_: mrdata, as its right up against the limit of my scope, I'm not sure anything is what its showing me
03:04 PM Tom_L: iirc wasn't it 100 ohm @ 0C or something like that
03:04 PM veverak: yup
03:05 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/avr/Toasteroven/
03:05 PM Tom_L: inflex wrote that
03:06 PM mrdata: veverak, is that a heating coil? resistance of those increases wih temperature iirc
03:06 PM veverak: it's temp. sensor
03:06 PM Tom_L: platinum sensor
03:06 PM veverak: it makes sense it resistance changes with temp ;)
03:07 PM mrdata: so, calibrate it by using a known temperature
03:07 PM veverak: I am not talking about calibrating
03:08 PM mrdata: voltage divider is easiest methinks
03:09 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/pdf/PT100_ref_data.pdf
03:09 PM Tom_L: there's the equasion
03:10 PM veverak: cool
03:10 PM Tom_L: we figured a 1.1v voltage reference was good enough for a toaster oven
03:10 PM Tom_L: the sensor is quite linear
03:11 PM rue_: I wonder if the latency on a cmos XOR gate is the same if its inverting as when its not
03:11 PM rue_: no that still wouldn't work
03:13 PM rue_: its almost like I'm looking for some kinda complimentrary line transciever
03:13 PM veverak: Tom_L: and used voltage divider?
03:14 PM Tom_L: directly connected to an avr pin
03:14 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/toaster_oven/control_sch.png
03:14 PM Tom_L: err maybe i lied
03:15 PM Tom_L: haven't looked at that in ages
03:16 PM Tom_L: there are some notes on that there
03:17 PM veverak: hmm, I see voltage divider with a cap
03:17 PM mrdata: veverak, what typical temperature range will use use this at?
03:19 PM veverak: room temp up to 300C
03:19 PM mrdata: typical analog input on an mcu will be voltage 256 levels from 0 to 5V i guess
03:20 PM veverak: got 16bit ADC i2c chip
03:20 PM veverak: :)
03:20 PM veverak: 3.3V
03:20 PM mrdata: so this sensor should give from 100 to 220 ohms or so, say
03:20 PM mrdata: 16bit? okay
03:21 PM veverak: 250 is 300c based on the table from the datasheet
03:21 PM veverak: 250ohm
03:21 PM mrdata: ok
03:22 PM mrdata: so if you have a stable voltage ref and a 500 to 1k ohm on top, you should get nice granularity
03:23 PM mrdata: from the voltsage divider
03:24 PM mrdata: how stable is your 3.3V
03:25 PM veverak: shared bus voltage I am afraid
03:25 PM mrdata: maybe you want around 1 V ref then
03:25 PM veverak: so, there is CAN bus with 5V, there is master device on CAN that converts that to 3.3V and uses i2c bus
03:26 PM veverak: and there will be the i2c 16bit analog, with 3.3V from that can device
03:26 PM veverak: I think that ref. diode is definetly a good idea
03:26 PM veverak: :)
03:26 PM veverak: which makes me think that "constant current source may be a better idea"
03:26 PM mrdata: why
03:27 PM veverak: hmm, yeah diode with voltage ref. is necessary anyway
03:28 PM mrdata: how much current are you happy with
03:28 PM veverak: do not have requirement
03:29 PM mrdata: so 2 mA is is okay
03:30 PM mrdata: so, a voltage divider with 500 ohms plus the sensor; fed by a 1 V ref
03:31 PM veverak: hmm a lot of people use amplifiers it seems
03:33 PM mrdata: should give 3300 or so values in the temperature range, so around 0.1 degree C precision
03:35 PM veverak: wow
03:35 PM veverak: fu this
03:35 PM veverak: they make IC's to measure these things
03:35 PM veverak: :)
03:37 PM veverak: MAX31865
03:39 PM rue_shop3: or buy a mercadies
03:39 PM veverak: ?
03:39 PM rue_shop3: rolls roice?
03:40 PM rue_shop3: about the same price as the average maxim chip isn't it?
03:40 PM rue_shop3: they do amazing things tho
03:41 PM rue_shop3: for example the max038 is better than what I been trying to build for a month now
03:41 PM veverak: it costs 4$
03:42 PM rue_shop3: in lots of?
03:42 PM rue_shop3: oo 74265
03:42 PM veverak: one piece
03:43 PM rue_shop3: well, if digikey sells it, thats only $12 ea
03:43 PM rue_shop3: ($8 shipping)
04:24 PM rue_shop4: wtf
04:25 PM rue_shop4: motorola took the 74xx175 and made a slower cmos version, the MC14175
04:25 PM rue_shop4: ?????
04:26 PM rue_shop4: I like how, while cmos was continiously re-inventing itself, ttl just moved forward
05:14 PM rue_shop4: hmm, a floppy drive probably has a bunch of ttl termination stuff in it, whats the data rate on a 1.4M floppy?
05:14 PM rue_shop4: drat, 500kbps
05:21 PM rue_shop3: this is all becuase the 1Mhz from my current fn gen is so ugly that I had to use a 555 to clean it up
05:21 PM rue_shop3: too bad a 555 cant do 10Mhz
05:45 PM rue_shop4: that said, at this point I can do one that can do a pretty nice 0.1-1Mhz
05:45 PM rue_shop4: with usb control
05:45 PM rue_shop4: just need to be able to get pulse shaping for 10Mhz speeds
05:49 PM rue_shop3: I suppose the nature of the problem is that,at those speeds, almost every little capacitance or inductance causes a response that throws the wave shape off
05:51 PM mrdata: yeah
05:51 PM mrdata: do you need nice curvy waves
05:51 PM mrdata: or rectangular
05:51 PM rue_shop3: well, I just want a descent square wave
05:51 PM rue_shop3: at 10Mhz, I'm not even pickey if its kinda round
05:52 PM rue_shop3: you should see the 1Mhz that comes off my HP gen, its horrid
05:52 PM mrdata: what defines decent? drift < 200 Hz?
05:53 PM rue_shop3: as far as wave shape, descent is a critically damped with a rise time of about 0.01us
05:54 PM rue_shop3: 100hz drift at 10Mhz would be a nice number, 10Hz would be awesome
05:54 PM rue_shop3: right now, the problem is the ringing on the final signal
05:54 PM rue_shop3: its about a 1V ring
05:54 PM rue_shop3: oh damn, I'll take pics
05:55 PM mrdata: RC low pass nails down the ringing
05:56 PM rue_shop4: well, I'll show you, final image is with 100pf load
05:57 PM rue_: as digital chips are all class AB, they put a woogle in the signal at every threshold
06:00 PM rue_: what your about to see is, all in all, pretty damn good
06:00 PM rue_: http://ruemohr.org/~ircjunk/projects/siggen/p1100337.jpg
06:01 PM rue_: this is the new board, its a 3 stage ring osc, where the voltage on the 4011 is varried to change its freq
06:01 PM mrdata: ok
06:01 PM rue_: the pot is the bias for the 74hc04 that squares up the result
06:01 PM rue_: http://ruemohr.org/~ircjunk/projects/siggen/p1100338.jpg
06:01 PM rue_: 1Mhz, 2.78V
06:02 PM rue_: http://ruemohr.org/~ircjunk/projects/siggen/p1100339.jpg
06:02 PM rue_: 10Mhz, 3.54V
06:02 PM rue_: evil ringing
06:02 PM rue_: http://ruemohr.org/~ircjunk/projects/siggen/p1100340.jpg
06:02 PM rue_: wholy &%$^#%%# THATS range! 3.8Khz, 1.55V, any lower than that and it goes random
06:03 PM mrdata: so you said 220pF eliminates the ringing
06:03 PM rue_: not completely
06:04 PM rue_: and it rounds things a LOT
06:04 PM rue_: http://ruemohr.org/~ircjunk/projects/siggen/p1100342.jpg
06:04 PM rue_: this is a redo, with a 74hc00 instead of a 4011
06:04 PM rue_: er, iirc
06:04 PM rue_: never-theless
06:04 PM rue_: http://ruemohr.org/~ircjunk/projects/siggen/p1100344.jpg
06:05 PM rue_: THIS is where I'm at
06:05 PM mrdata: so put a series R in there with the 220p
06:05 PM rue_: this is on a solderless breadbaord, thats 10Mhz
06:05 PM rue_: that lower ring is a problem
06:05 PM rue_: thats the signal off the maxim rs422 reciever
06:06 PM rue_: thats 0.1us/div
06:06 PM rue_: and thats with 100pf load to grond
06:08 PM mrdata: so try 10 pF and 1 ohm in series, to ground
06:08 PM rue_: 10pf isn't enough, the ringing all gets louder
06:08 PM rue_: tho I wonder, when Iwas driving the power amp last night, some of the ring didn't come thru
06:09 PM mrdata: the ringing gets louder because tau is too small; the R increases tau
06:09 PM mrdata: try R up to 100 ohm
06:09 PM rue_shop4: parallel?
06:11 PM mrdata: series
06:11 PM rue_shop4: 10R in series w/100pf is a TINY bit better
06:12 PM mrdata: increase R
06:13 PM rue_shop4: I think I need to make a decade box, this bench is getting insane, where did the 1R go?
06:14 PM mrdata: please try 100pF and 100 ohm R in series
06:15 PM rue_shop4: cant find the 100R... give me a min :)
06:16 PM rue_shop4: oh 1% there it is
06:17 PM rue_shop4: from that last image, the bottom ring is a bit louder, and the top ring matches the bottom one
06:17 PM rue_shop4: w/ 100R, 100pf
06:17 PM rue_shop4: series
06:19 PM mrdata: but 10 ohm was a teeny bit better?
06:19 PM mrdata: i think the sweet spot is somewhere from 10 to 100 ohm and 10 to 100 pF
06:20 PM rue_shop4: oh crap, I just blew my high speed power amp
06:20 PM rue_shop4: accidently shorted it to +12
06:20 PM mrdata: but it might be somewhere from 10 to 200 ohm and 10 to 200pF
06:20 PM mrdata: wtf
06:21 PM rue_shop4: wonder if I have another LM2419
06:22 PM rue_shop4: ARG, that hurts
06:25 PM rue_shop4: awe, this is why I hate having to scrounge parts
06:25 PM rue_shop4: might take me another day of just searching to find antoher one
06:25 PM rue_shop4: it was the best line driver I could find
06:27 PM rue_shop4: I could give up and only use this from 0.1 to 1Mhz
06:27 PM rue_shop4: use a 555 as a line driver
06:28 PM rue_shop4: in the meantime, the soldering pen heated up and I have the 260pf variable cap
06:29 PM rue_shop4: that was a damn nice driver
06:29 PM rue_shop4: ok, 500R variable and a 260pF variable, lets find waldo
06:33 PM rue_shop4: almost 0R and 260pF
06:36 PM rue_shop4: 11.6R and 252pF
06:36 PM rue_shop4: is best
06:36 PM rue_shop4: does it simulate into anything reasonable?
06:36 PM rue_shop4: at 10Mhz. I'm gonna dial it down and see what it looks like at lower freqs
06:46 PM rue_shop4: so, hmm
06:47 PM rue_shop4: maybe what I'll do, is make up a little adjustable terminator board, with a 500R and an adjustable cap
06:47 PM rue_shop4: need to find more radio caps
06:52 PM rue_shop4: hmm
06:52 PM rue_shop4: mrdata, you have a pretty good feel for this stuff :)
06:55 PM rue_shop4: so, the osc is a 74hc00 centred on a variable voltage, its cleaned up by a 74hc04 now @ 3.3V... er, apparently my white led shunt regulator is 2.5V... and then thresholded by a MAX3468 ($$)
06:55 PM rue_shop4: so, output driver is one of the next problems, being I just blew up my LM2419
06:56 PM rue_shop4: I need to make a brick wall so I can bang my head on it when stuff like that happens
06:57 PM rue_shop4: I have 7x 7490 (no, no LS or nothin, just origional 7490 from the stone age) divide-by-10 ready
06:57 PM rue_shop4: a 74ls151 to select which divider tap to use
06:57 PM rue_shop4: 10 switches specially reserved for a good project
06:57 PM rue_shop4: a BNC connector
06:58 PM rue_shop4: this project needs a parts box
07:02 PM mrdata: 11.6R and 252pF is 54 MHz
07:04 PM mrdata: f = 1 / ( 2 pi RC )
07:04 PM rue_shop4: I really should sit down and write that command line calc for that
07:04 PM rue_shop4: so, 54Mhz @-3db isn't bad for a 10Mhz signal
07:05 PM mrdata: yeah
07:05 PM rue_shop4: its not a really pretty shape, but its as good as it gets
07:05 PM rue_shop4: oh picture!
07:07 PM rue_shop4: the rise time sucks, fall time is descent
07:07 PM mrdata: this filter shouldnt harm lower frequencies; the ringing relates to the rise time
07:07 PM rue_shop4: yes, I dialed it down and it all looks good
07:08 PM rue_shop4: I will play more, I'm wondering if I do some filtering at earlier stages of the whole thing
07:08 PM rue_shop4: see if I can clean things up the whole way thru
07:09 PM rue_shop4: oh, AND this is the breadboard ver
07:11 PM rue_: http://ruemohr.org/~ircjunk/projects/siggen/p1100345.jpg
07:12 PM mrdata: ok
07:12 PM rue_: maybe if i can clean the signal up right back to within the stages of the ring osc, I can get that rise time fixed up
07:12 PM rue_: http://ruemohr.org/~ircjunk/projects/siggen/p1100346.jpg
07:12 PM rue_: note the 2.5V shunt regulator made of a white LED ;)
07:13 PM mrdata: heh
07:13 PM rue_: I'll do a schematic while its still working descently
07:14 PM rue_: tho I'm surprised, I thought LEDs generated diode noise, didn't notice anything
07:15 PM Tom_L: just scored a nice micrometer free
07:16 PM rue_shop3: cool
07:16 PM Tom_L: https://www.amazon.com/Mitutoyo-Digimatic-MDC-MX-Outside-Micrometer/dp/B00WML73OG
07:16 PM rue_shop3: oh, and apparently, doing a ring osc with 74xx00 will be slow, should be faster with a 74xx02
07:16 PM Tom_L: brand new in the box in shrinkwrap
07:17 PM rue_shop3: wow
07:17 PM Tom_L: my kid won it at the national and he already has a nice one
07:17 PM Tom_L: nice n accurate
07:17 PM Tom_L: 5 place decimals
07:19 PM Tom_L: rather heavy for a mic..
07:19 PM Tom_L: won't warp as much
07:20 PM Tom_L: i think i'll keep that one in my desk, not in the shop
07:23 PM Tom_L: he got a nice 6" calipers too but he kept those, much nicer than his first chinese one
07:36 PM rue_: http://ruemohr.org/~ircjunk/projects/siggen/p1100348.jpg
07:37 PM rue_: mrdata, is that enough detail for ya?
07:37 PM rue_: might have to bump the size
07:37 PM mrdata: cool
07:38 PM rue_: just resposted maybe a bit larger
07:38 PM rue_: damn pencil dosn't come up well
07:39 PM rue_: if I put some termination between the gates of the ring it might help
07:39 PM mrdata: so, what's going on with the second half there
07:39 PM rue_: dunno if that'll mess with the freq
07:39 PM rue_: its the level conversion and cleanup
07:39 PM mrdata: oh, because the oscillator doesnt output TTL levels
07:40 PM rue_: the osc at the top I balanced, so its 0 is consistant
07:40 PM rue_: so, looking for 2.5V crossings should be consistant
07:40 PM rue_: but I dont have any comparators fast enough
07:41 PM rue_: and stuffing a diff signal right from the last gate into the max3468 didn't seem to be happy
07:41 PM rue_: tho I'm tempted to try again
07:42 PM rue_: I think the duty comes out badly cause of the propigation delay on thelast gate, with that bias I can tweek it
08:48 PM rue_shop4: I think the 4011 worked better just cause its a slower chip
08:53 PM rue_shop4: on average, the stages need 24 ohms and 100pf, but its dragging the osc down a bit too much
08:53 PM rue_shop4: I'm gonna play with the cmos oscs in the new setup with the new output
09:04 PM rue_shop4: huh, ok, motorola cmos chips suck
09:33 PM rue_shop4: 15mins later... "DAMN, THE SIGNAL SURE CLEANS UP IF YOU RECONNECT THE SCOPES GROUND LEAD"
09:40 PM rue_shop4: ok, the max3468 doesn't work with the 4011 using 1 gate as an inverter, at 10Mhz the propigation delay is almost as long as the cycle time
09:40 PM rue_shop4: but
09:51 PM rue_shop4: the funny thing with the cmos version is that ONE of the links needs to be made with a resistor
09:52 PM rue_shop4: the resistor needs to be more than 1M, but can be up to 10M
09:52 PM rue_shop4: and it dosn't make a slight bit of difference to it
09:52 PM rue_shop4: !??!?!?!
10:26 PM rue_shop4: huh, the 4011 with the 74hc04, and the new ring snubber, is the best yet
10:34 PM rue_: one of the advantages of time travel is that the laundry gets finished really fast
10:52 PM rue_shop4: disadvantage of course being how fast a vaccation ends
10:52 PM rue_shop4: "What do you mean I'm back 3 years late!?"
11:28 PM rue_shop4: mmm the 10Mhz from the current circuit looks better than the 1Mhz from the signal gen
11:30 PM rue_shop4: slew at 4.5V/0.1us
11:30 PM rue_shop4: hmm
11:40 PM Tom_L: is it wrong to make a lathe part on the mill if i know i can make it more accurately than on the manual lathe?
11:51 PM mrdata: no
11:52 PM mrdata: if you need accuracy, go for the accuracy