Resistor in reset line?

Thread Starter

metermannd

Joined Oct 25, 2020
189
Looking again at one board I have...

There's one design bit that has me scratching my head.
There's a 100-ohm resistor smack dab in the middle of the reset line.
I would like to see if the brain trust can give me any insight as to why...

On one side of the resistor, I have one logic gate (input end), the CPU (similar to MC6801), and the connection to the deadman circuit on another board.

On the other side of the resistor is the MC6840 timer and MC6821 PIA.
 

Delta Prime

Joined Nov 15, 2019
996
Hello there :) I am Sergeant at arms for the Brain trust after consulting with the Brain trust they have determined that you are not worthy,.. just kidding. But they did mention if you would happen to have a schematic diagram or a photo an etch a Sketch to draw what you are seeing. Thank you the Brain trust has spoken,
 

nsaspook

Joined Aug 27, 2009
8,382
Usually the series resistor is a damper to reduce ringing at the pin on the CPU.


https://resources.altium.com/p/damping-and-reflection-transfer-series-termination-resistor
There are two issues to consider when sizing a series termination resistor for use with a particular driver in your PCB:

  • Suppressing reflection: the series resistor at the source can suppress reflection at the source back into the driver. Similarly, termination at the receiver end will suppress reflection back along the transmission line.
  • Damping: a series termination resistor will increase the damping constant in the equivalent RLC circuit model for a transmission line. If the series termination resistor takes just the right value, you can critically dampen the transient oscillation that occurs due to switching and suppress overshoot, albeit at the cost of increasing the rise time.
 

MrChips

Joined Oct 2, 2009
23,515
+1
Agreed.
Series resistance is to suppress reflections back to the driver.
Parallel resistance at the load is to terminate the transmission line which in turn suppresses reflections.

1613311772891.png
 

Thread Starter

metermannd

Joined Oct 25, 2020
189
Delta: I'm working on an unified schematic of all the boards on one page, it may be a few days.

nsaspook, MrChips: Thank you. Just wanted to make sure it wasn't some kind of kludge.
 

DickCappels

Joined Aug 21, 2008
7,426
A reset input probably would not require damping of ringing as say a clock line might need, but maybe engineer was just being overly cautious.
 

AnalogKid

Joined Aug 1, 2013
9,248
Keep hoping.

After too long I found a resistor wandering around in the middle of a bunch of disconnected stuff. But other than two capacitors that serve only to complicate things more, it is not connected to anything. Holy Data General, Batman.

<deleted micro-rant>

1. Just so we're all literally on the same page -- what are the sheet name and reference designator for the resistor we are discussing?

2. Can you at least give us a list of the reference designators of the parts that touch the reset signal?

ak
 
Last edited:

Thread Starter

metermannd

Joined Oct 25, 2020
189
Kinda figured I'd get a bit of grief for the 'shattered' schematic layout. Happened for a bunch of reasons - analysis, removing unused options, replacing unavailable parts, etc. I'll have to start pasting everything back together soon.

Here are the relevant sections in two closeups: (which I should have posted instead of that merged schematic; KiCAD changed the designators on me when I combined the schematics!)

Logic.pdf:
Reset line (J3 pin 4, U8 pin 40, R4, C12); Res2 line (R4, C13, U7 pin 8, and U10 pin 37).

Analog.pdf:
Reset line apparently originates from U3B and D11 / D12 junction.

Additional notes - U8 is a variant of the 6801MCU, and it appears the relevant section of the Analog board may be a watchdog circuit.
 

Attachments

MrChips

Joined Oct 2, 2009
23,515
Agreed. The purpose of an RC delay would be to hold some circuits in a RESET state until other circuits get a chance to initialize.
 

Thread Starter

metermannd

Joined Oct 25, 2020
189
Coming back to this... the 8279 display / keyboard controller has a reset pin, which is currently fed from an independent circuit, consisting of a 10K / 0.01uF cap (giving a 100uS delay), fed through an inverter to give the needed pulse.

I believe the reason it wasn't tied into the reset line on the logic board was that there was no room to carry the reset line over through the retrofit connection between the logic and I/O boards.

I now happen to have one unassigned position on the inter-board connector between the I/O and logic boards.

Should I bring a connection from the reset line into the I/O board, and if so, from which side should I take the signal?
Reset or Res2?
 

AnalogKid

Joined Aug 1, 2013
9,248
It ain't broke. Don't fix it.

Also, R9-C14 is not a delay network per se. It sets the reset pulse width. At power-on, U24B pin4 goes high immediately, and stays high until C14 charges up to approx 0.5 x Vcc, or 0.7 x R x C.

Consider replacing the 74xx04 with a 74xx14. This gate has hysteresis that handles relatively slow edges better than a standard input. Granted that a 100 us time constant should be fast enough not to cause problems, and it is a reset input, not a counter or something else where a noise burst would be a real problem. But it is good practice to condition R-C edges with a hysteretic gate input. Not that the resulting pulse width will decrease by approx. 10%.

ak
 
Top