IC1 74AC574 issue.
Posted: Tue Feb 21, 2012 1:54 pm
A problem on a local DigiLite has highlighted an issue with IC1, the 20 pin chip on the underside of the pcb. This is an 8 bit latch / driver. It may well be just one chip, but it's worth mentioning the implications. If your DigiLite is working, then you almost certainly don't have a problem.
IC1 latches the digital I/Q signals from the dsPIC and acts as a driver into the modulator section. On this chip, I and Q each have 3 inputs connected together and the corresponding 3 outputs are connected together to increase the drive. The problem unit was showing lots of noise on the power supply rails and on the spectrum analyser. It turned out to be caused by not all of the 3 tied outputs changing at the same time, which meant that a large amount of current was being drawn because of an output of 5v being connected to an output of 0v.
As the design has progressed from "Poor Man's DATV" to DigiLite, various parts of the circuit have been run at various voltages. On DigiLite, the digital section runs at 3.3v and the modulator section at 5v. The 74AC574 has a 5v supply, so each output is at either at 5v (high) or 0v (low), depending on the input. The inputs are driven by the outputs of the dsPIC, which are either at 3.3v (high) or 0v (low). The minimum guaranteed input threshold of an AC574 to give a high output is 3.5v, so the high from the dsPIC is below the threshold. In this case, it's possible that the 3.3v from the dsPIC was right on the threshold and it was high enough for some of the 3 inputs, but not all of them.
Chips are designed to work over a wide temperature range, so there is usually quite a bit of leeway. A better chip to use for IC1 is the 74ACT574. The 'T' version has an input threshold of 2v, so the 3.3v output from the dsPIC is well above it.
You can check for the problem if you have a scope. Locate R9 and R10, south west of the I level pot. On the left side of these are the I and Q outputs from the 574. Check that the squarish wave levels are either 5v or 0v, with no other steps visible in between. In Phase test mode should make it easier to see. If there are no steps, then the problem is not there.
To summarise, this may be an isolated incident with one chip. if your DigiLite is working then there is almost certainly no problem. If the problem is there, or you haven't yet bought a 74AC574, use a 74ACT574. e.g. Farnell 959-0412. I'm now using one of these on my DigiLite.
Brian
IC1 latches the digital I/Q signals from the dsPIC and acts as a driver into the modulator section. On this chip, I and Q each have 3 inputs connected together and the corresponding 3 outputs are connected together to increase the drive. The problem unit was showing lots of noise on the power supply rails and on the spectrum analyser. It turned out to be caused by not all of the 3 tied outputs changing at the same time, which meant that a large amount of current was being drawn because of an output of 5v being connected to an output of 0v.
As the design has progressed from "Poor Man's DATV" to DigiLite, various parts of the circuit have been run at various voltages. On DigiLite, the digital section runs at 3.3v and the modulator section at 5v. The 74AC574 has a 5v supply, so each output is at either at 5v (high) or 0v (low), depending on the input. The inputs are driven by the outputs of the dsPIC, which are either at 3.3v (high) or 0v (low). The minimum guaranteed input threshold of an AC574 to give a high output is 3.5v, so the high from the dsPIC is below the threshold. In this case, it's possible that the 3.3v from the dsPIC was right on the threshold and it was high enough for some of the 3 inputs, but not all of them.
Chips are designed to work over a wide temperature range, so there is usually quite a bit of leeway. A better chip to use for IC1 is the 74ACT574. The 'T' version has an input threshold of 2v, so the 3.3v output from the dsPIC is well above it.
You can check for the problem if you have a scope. Locate R9 and R10, south west of the I level pot. On the left side of these are the I and Q outputs from the 574. Check that the squarish wave levels are either 5v or 0v, with no other steps visible in between. In Phase test mode should make it easier to see. If there are no steps, then the problem is not there.
To summarise, this may be an isolated incident with one chip. if your DigiLite is working then there is almost certainly no problem. If the problem is there, or you haven't yet bought a 74AC574, use a 74ACT574. e.g. Farnell 959-0412. I'm now using one of these on my DigiLite.
Brian
I'd forgotten that I ordered two at the same time, probably from Farnell. It looks like a TI chip and the code on it is: 04ALR8K. Chances are that my original, which showed no problems, has the same markings, but I don't have it any more.