Enable Pin G and G (Complement) on am26ls32ac

Thread Starter

eb123

Joined Jul 3, 2017
74
The am26ls32ac (http://www.ti.com/lit/ds/symlink/am26ls32ac.pdf) differential receiver IC includes two pins, G and G (Complement) which appear to have some effect on the output of each receiver.

The description on page 11 is not entirely clear to me, so I'll just explain what I need:

Two receivers will be used to process a TTL encoder's differential input. The output, which should always be enabled, should provide TTL-logic level outputs to the receiving MCU. Can I just provide G with 5v, and connect G (Complement) to ground? Do the output pins need to be connected to G/G(complement)? What exactly is the purpose - a few example may help illustrate this?
 

ebp

Joined Feb 8, 2018
2,332
See Table 1 of the datasheet.

Those inputs are simply "tri-state" control for the outputs of the receiver. If neither input is asserted the outputs are set to high impedance. Because it is an OR gate, it is sufficient to assert either input, though it is generally not good practice to leave unused inputs floating (not really terrible with TTL if the logic level truly doesn't matter, but a definite no-no for any CMOS family). You can do this as you suggested or simply by tying both together and to either 5 V or ground . Do what is most convenient from a layout perspective.
 

MrChips

Joined Oct 2, 2009
30,714
As the man says, use whichever pin matches the logic of your control signal.
Wire /G to high and use G for positive logic control.
Wire G to low and use /G for negative logic control.
 

Thread Starter

eb123

Joined Jul 3, 2017
74
As the man says, use whichever pin matches the logic of your control signal.
Wire /G to high and use G for positive logic control.
Wire G to low and use /G for negative logic control.
Could you please elaborate on what you mean by positive/negative logic control?
 

Thread Starter

eb123

Joined Jul 3, 2017
74
See Table 1 of the datasheet.

Those inputs are simply "tri-state" control for the outputs of the receiver. If neither input is asserted the outputs are set to high impedance. Because it is an OR gate, it is sufficient to assert either input, though it is generally not good practice to leave unused inputs floating (not really terrible with TTL if the logic level truly doesn't matter, but a definite no-no for any CMOS family). You can do this as you suggested or simply by tying both together and to either 5 V or ground . Do what is most convenient from a layout perspective.
Is there any difference in output between my suggestion and yours? Am I still guaranteed TTL-logic level outputs (i.e. Low:0-0.4,High:2.7+)?
I'm still not sure what the pins actually do, and why they are needed.
 

ebp

Joined Feb 8, 2018
2,332
So-called "tri-state" (actually is or was a registered trademark of ??) outputs either behave exactly like a normal output for whatever logic familiy is in question when enabled or are high-impedance (essentially, for practical purposes, open-circuit) when disabled - The purpose of using tri-state devices is to allow multiple outputs to drive a single data line. All are kept disabled except the one which is permitted to control the line at any particular time. It is very extensively used with data buses.

An OR gate with one active-low and one active-high input is provided. Sometimes you only get a single control input. The originators of the part in question (AMD for the original standard TTL type) persumably just decided that since there would otherwise be an unused pin on the package, they might as well give users the option of active-high or active-low enable, just for convenience (e.g. if you had two receivers sharing an output line you could enable one active-high, the other active-low with the same control signal). It makes absolutely no difference to how the output behaves which you choose to use.

You should be able to find all sorts of info on this topic on the web.
 
Last edited:

MrChips

Joined Oct 2, 2009
30,714
Could you please elaborate on what you mean by positive/negative logic control?
Positive logic means that you need to send 1 or logic-high to activate the function.
Negative logic means that you need to send 0 or logic-low to activate the function.

This is what the bar over the G means. The over-score over the symbol is to remind you that the control signal confirms to negative logic.
Hence, G means that the signal requires positive logic.
/G means that the signal requires negative logic.
 

Thread Starter

eb123

Joined Jul 3, 2017
74
The table also shows a Low output
So-called "tri-state" (actually is or was a registered trademark of ??) outputs either behave exactly like a normal output for whatever logic familiy is in question when enabled or are high-impedance (essentially, for practical purposes, open-circuit) when disabled - The purpose of using tri-state devices is to allow multiple outputs to drive a single data line. All are kept disabled except the one which is permitted to control the line at any particular time. It is very extensively used with data buses.

An OR gate with one active-low and one active-high input is provided. Sometimes you only get a single control input. The originators of the part in question (AMD for the original standard TTL type) persumably just decided that since there would otherwise be an unused pin on the package, they might as well give uses the option of active-high or active-low enable, just for convenience (e.g. if you had two receivers sharing an output line you could enable one active-high, the other active-low with the same control signal). It makes absolutely no difference to how the output behaves which you choose to use.

You should be able to find all sorts of info on this topic on the web.
So, active-high means that for the pin to be enabled, the input must be high (5v in this case I think), and active-low means that for the pin to be enabled, input must be low (GND)?

The table shows Hight and Low outputs under different conditions. I'm unable to understand these conditions, but I presume they mean output is high when the decoded signal is high, and low when the decoded signal is low?
 

MrChips

Joined Oct 2, 2009
30,714
Connect G to high OR /G to low and your AM26LS32 will be in working mode.
Set G to low AND /G to high and your AM26LS32 will stop working.
 

Thread Starter

eb123

Joined Jul 3, 2017
74
I've finally gotten around to soldering the component, and testing it out.
The datasheet specifies a transfer rate of 52Mbps. Is this to say, 52Mbps for each receiver (there are 4)?

If so, this seems like quite a lot, and I would assume that the IC is capable of handling a high speed encoder?

On a separate note, I've noticed that the device itself is rather old (1990). Are devices this old still used in new circuit designs (professionally), or do electronic engineers tend to discriminate against the older ICs in favour of the newer ones? I personally don't see why it would matter, as long as it does the job.
 

ebp

Joined Feb 8, 2018
2,332
Yes, the bit rate is per receiver. I'm not sure I'd want to be anywhere within several kilometres of a rotary encoder on something spinning fast enough to get anywhere near that rate ;) I certainly wouldn't want to have to pay for the bearings.

I used the original (non-LS) version of this part sometime well back in the previous millennium. Often old parts do fall out of favor because there are newer parts that have better performance one way or another. You can probably find lots of receivers with lower "quiescent" power consumption - that is when the input is steady-state. There are also newer receivers that have built-in control of output behavior (usually called "fail-safe" operation) if the input is open circuit or short circuited, which is highly desirable in some applications. I'm not entirely certain, but I think there may be some receivers that have deliberately-reduced frequency response to make them more noise immune.

With older parts it is worth checking to see if there are still at least two manufacturers producing them so you don't suddenly find yourself without a source. Advanced Micro Devices (AMD), who first introduced the part, quit making them many years ago. The other alternative is a "pin compatible" part which will function as well or better and has the same pinout. These days you'll sometimes find an ancient part is now being made only in a surface mount package.

Some old parts are still very popular. The 78xx series three terminal regulators have been around much longer than the 26LS32 and are still hugely popular. I'm surprise how many people look for help on AAC with circuits using the 741 op amp, which is older still.
 
Top