Don't forget the pull-up/down !

[sporniket]

Outside of retro stuff, I do some basic electronics for my daughter, and tinker with a microcontroller.

Until recently, I was blaming the perfboard assembly or fingers touching the solder joints for the weird behaviour of signals triggered by push buttons (like unwanted trigger when moving the perfboard holding the buttons or button releases not registered).

But in fact, thinking about the mandatory fixes established by @exxos, it suddenly struck me that I forgot to add some pull-down . A little test later with a spare resistor on one of my latest montage, and it works flawlessly...

mandatory fixes win again... and not only for ST computers

[Steve]

Is there a tutorial or explanation for the pull up/pull down resistors? I would like to know how they work in relation to the ST address and data bus, how this effects the chips, the power ... Could someone explain in simple terms?

- When dropping from 10k SILS on the bus to 2.2k, what is happening electrically? Is it something to do with raising/lowering TTL values so that noise doesn't accidentally trigger TTL or something?

Thanks,

[stephen_usher]

Think of the wire as a trough of water.

The pull-ups refill the trough to a certain level. The lower the resistance, the faster the "pull-up" can refill the trough after another device has emptied the water out. The longer the trough, the more water there is to refill. and the longer it takes for the water to flow to the other end.

There is a balance, however. You don't want to be able to pump in more water than the devices can drain at any one time. Also, to stretch the analogy further, you also don't want to overload the pumps on the devices and cause them to break either. (In this case, TTL chips are shorting the wire to ground. If the pull-up resistor is too low the current flowing can overheat and damage the internal transistor in the chip.)

The TTL specification gives a maximum current loading and the recommended resistance of a pull-up resistor for TTL is hence in the region of 10K ohm. Lowering the resistance below this is technically out of spec. and could potentially damage components but if the capacitance of the trace is large enough you may need to do so to allow the transitions from HIGH to LOW to be fast enough.

[Steve]

Think of the wire as a trough of water.

The pull-ups refill the trough to a certain level. The lower the resistance, the faster the "pull-up" can refill the trough after another device has emptied the water out. The longer the trough, the more water there is to refill. and the longer it takes for the water to flow to the other end.

There is a balance, however. You don't want to be able to pump in more water than the devices can drain at any one time. Also, to stretch the analogy further, you also don't want to overload the pumps on the devices and cause them to break either. (In this case, TTL chips are shorting the wire to ground. If the pull-up resistor is too low the current flowing can overheat and damage the internal transistor in the chip.)

The TTL specification gives a maximum current loading and the recommended resistance of a pull-up resistor for TTL is hence in the region of 10K ohm. Lowering the resistance below this is technically out of spec. and could potentially damage components but if the capacitance of the trace is large enough you may need to do so to allow the transitions from HIGH to LOW to be fast enough.

That is very intersting and a great analogy thank you so much I have been expermimenting with the bus resistors on my TT lately because of some strange behaviours I've been experiencing. I socketed them last week so I'm going to run them with the stock 10ks, then drop to 4.7ks, then 2.2k and just see how it's behaviour changes.

[stephen_usher]

Looking at the massive under-shoot/overshoot and ringin on the cartridge port on my TT I'm thinking about adding some tiny capactors (~1pF) to the address/data lines to see if I can suppress the ringing. It would be nice to be able to have a break-out board which can fit into the cartridge port with a pass-through so as to connect a cartridge just to debug things. It's a pity that the edge connector is so non-standard.

[Steve]

Looking at the massive under-shoot/overshoot and ringin on the cartridge port on my TT I'm thinking about adding some tiny capactors (~1pF) to the address/data lines to see if I can suppress the ringing. It would be nice to be able to have a break-out board which can fit into the cartridge port with a pass-through so as to connect a cartridge just to debug things. It's a pity that the edge connector is so non-standard.

Would you suggest that:
Ideally: The TT runs stable with 10k SIP values
If possible: 4.7k
Last resort, but use them if you have to: 2.2k

Also the TT does not seem to have a large 4700uf cap like most Ataris - I wonder if fitting one somewhere would be beneficial.

[stephen_usher]

It's ringing on the individual data lines, so it needs capacitors on each line to dampen the oscillation. The ringing is enough that a fast device could see the line transition from LOW to HIGH, back to LOW and up again.