Pis and serial ports...

I'm guessing that the trend is to more "internal" hubs, not fewer.

So since the enumeration limit is software-imposed, should we expect to see the limit raised for Raspbian?

--
-michael - NadaNet 3.1 and AppleCrate II: http://home.comcast.net/~mjmahon

The Pi has worked for me with a USB dongle containing a SIM which reports 3.6 Gb/s connect speeds and gives download rates of about 250 KBytes/sec over what is seen by 'wvdial' (or was it kppp?) as a serial link. Used this for a month or so, a couple of months ago.

And I've used USB-to-serial converters with other Linux distros on other PCs to drive old-style RS232 external serial modems, so I'd expect that the same _ought_ to work on a Pi.

But of course YMMV.

--
Windmill, TiltNot@NoneHome.com       Use  t m i l l 
J.R.R. Tolkien:-                            @ S c o t s h o m e . c o m 
All that is gold does not glister / Not all who wander are lost

Any USB 2 or USB 3 'hub' is actually a 'switch' with store and forward and dealing with different timing constraints for upstream or downstream being slower. As well as recording details of device bus type for itself and upstream of downstream connections.

Often they have blocks inside the device called a Transaction Translater (or similiar) to deal with this.

Having had to deal with this in past and designing kit with its own USB 2 'hub' inside the product.

--
Paul Carpenter          | paul@pcserviceselectronics.co.uk 
    PC Services 
  Raspberry Pi Add-ons 
 Timing Diagram Font 
  GNU H8 - compiler & Renesas H8/H8S/H8 Tiny 
 For those web sites you hate

From what I've read, all hubs (yes, I'll use that word because that is what they are called, even if it is not a good description) have a TT. Cheap and nasty hubs have only one. Better hubs have one per port. You can see this on the output of lsusb -t, it will show "single TT" or "TT per port". I think (not an expert in this field by a long way) that single TT devices can't handle different speeds (USB1.x/USB2) on different ports and default to the slowest. I think there is also some issues with transaction queueing on some devices like USB audio input/output.

A USB hub behaves like an Ethernet switch in that it stores and forwards.

Dave

That's not all a switch does though.

I meant, of course, 3.6 *megabits* per second, not giga !

--
Windmill, TiltNot@NoneHome.com       Use  t m i l l 
J.R.R. Tolkien:-                            @ S c o t s h o m e . c o m 
All that is gold does not glister / Not all who wander are lost

to

theory,

Modern stuff maybe, I wouldn't be so sure aboout old kit, old being

20+ years. Things tended to stick to the RS232 specs a bit tighter than they do now. If things were happy at TTL levels Maxim wouldn't have needed to develop MAX232 etc range of chips... B-)

--
Cheers 
Dave.

Actually even the 1488/1489 drivers, which are what was used in the early eighties, work OK with TTL levels.

The MAX232 was not developed because it was required to make things work, but because it was required to deliver equipment that adheres to standards. That is a different thing.

Modern stuff, like everything since 1980. see "MC1489".

The earliest reference to that part I can find is a 1975 Harris Semiconductor catalogue, but that's a Motorola part-number so It dates from before that.

--
?? 100% natural

yes, but groping into archived storage in the memory cells ISTR that such chips were often driven indirectly, using a 'balanced line driver' chipset.

i.e. it was not possible to integrate the logic of decoding and bufferning a serial bitstream with the problems of driving a +- 12v line or detecting such and turning it into a 5v logic signal.

--
Ineptocracy 

(in-ep-toc?-ra-cy) ? a system of government where the least capable to lead are elected by the least capable of producing, and where the members of society least likely to sustain themselves or succeed, are rewarded with goods and services paid for by the confiscated wealth of a diminishing number of producers.

There is a *lot* of older industrial kit that needs the full +- swing. There is also some that 'works mostly' with TTL levels and as a result an absolute nightmare to fault-find :(

--
W J G

Nor is it even how all ethernet switches operate. "Cut-through" switches are popular in some environments where low latency is paramount. They are still true switches. Store-and-forward is only one way a switch can be implemented, it is not the definitive implementation by any stretch of the imagination.

--
Andrew Smallshaw 
andrews@sdf.lonestar.org

Well, a USB to RS-232 dongle *IS* a ready made solution, also. It's not like there is a kit or something.

Obviously. But plenty of people seem to have knocked together their own solution to use the R-Pi UART, as an alternative to using a ready-made product.

It is the old stuff I am referring to. If you think about how the receiver would be implemented, you will see that it is just a standard input. The whole bipolar, wide voltage swing is to deal with external noise issues. So that with a lot of noise on the input, you still get a transition at the threshold voltage of the receiver input.

Maxim didn't invent RS-232, in fact, I wouldn't be surprised if Maxim wasn't even around when RS-232 was first used.

--

Rick

wiki:

The RS-232 standard defines the voltage levels that correspond to logical one and logical zero levels for the data transmission and the control signal lines. Valid signals are either in the range of +3 to +15 volts or the range ?3 to ?15 volts with respect to the ground/common pin; consequently, the range between ?3 to +3 volts is not a valid RS-232 level. For data transmission lines (TxD, RxD and their secondary channel equivalents) logic one is defined as a negative voltage, the signal condition is called "mark". Logic zero is positive and the signal condition is termed "space". Control signals have the opposite polarity: the asserted or active state is positive voltage and the deasserted or inactive state is negative voltage. Examples of control lines include request to send (RTS), clear to send (CTS), data terminal ready (DTR), and data set ready (DSR).

The standard specifies a maximum open-circuit voltage of 25 volts: signal levels of ±5 V, ±10 V, ±12 V, and ±15 V are all commonly seen depending on the voltages available to the line driver circuit. Some RS-232 driver chips have inbuilt circuitry to produce the required voltages from a 3 or 5 volt supply. RS-232 drivers and receivers must be able to withstand indefinite short circuit to ground or to any voltage level up to ±25 volts. The slew rate , or how fast the signal changes between levels, is also controlled.

Because the voltage levels are higher than logic levels typically used by integrated circuits, special intervening driver circuits are required to translate logic levels. These also protect the device's internal circuitry from short circuits or transients that may appear on the RS-232 interface, and provide sufficient current to comply with the slew rate requirements for data transmission. ==========================================

so 0 to 5VDC is simply not going to work. a minimum of -3 to plus 3v is needed to meet the spec. the fact that UARTs work on 0-5V inputs is irrelevant: they dont get directly connected to the lines.

--
Ineptocracy 

(in-ep-toc?-ra-cy) ? a system of government where the least capable to lead are elected by the least capable of producing, and where the members of society least likely to sustain themselves or succeed, are rewarded with goods and services paid for by the confiscated wealth of a diminishing number of producers.

I would like to know the receiver they are using that doesn't have a specific threshold voltage, but requires a wide input swing.

The "range" of voltage in an input is a spec issue. In TTL there is a range specified of 0.8 to 2.0 volts at the input, depending on the logic family. The output spec is 0.4 to 2.8 volts for most TTL families. That doesn't mean you have to drive it to those voltages. Those specs give reliable operation in the presence of noise. The input is guaranteed to be seen as a '0' if the voltage is below 0.8 volts and the output guarantees a driving voltage of 0.8 volts, so you can have up to

0.4 volts of noise and still have reliable operation. Likewise with the high level voltage.

But the TTL inputs all still switch at about 1.1 volts give or take with changes in temperature. That's the point. The receiver has a threshold at a single point which may vary a bit. The rest of the spec is about guarantying operation in the presence of noise. If you know you have little or no noise, you can operate with a smaller margin than the spec.

RS-232 is the same. The unipolar receivers have a single input threshold voltage at about 1.0 volts. If you drive this input with a low noise TTL driver it will reliably cross the input threshold. The polarity will be reversed however, so you still need to account for that.

It is *possible* that a differential receiver (RS-422 type) is being used to receive an RS-232 signal. Then they would use the cable ground as the reference point and driving with TTL would not work well if at all. But this is very unlikely and would only exist in high end equipment that was rather over designed.

--

Rick

Yes of course the receiver has a single (poorly defined) switching point, but that's no use at all if the transmitter doesn't send a high enough swing to overcome noise.

In a factory environment, even shielded cable can easily end up with a couple of volts of noise - especially in places I've been where there are over a hundred yards of cable. Have you seen just how long a 5 ply corrugated cardboard maker is? Or how far away poly granules are siloed from the moulding machines?

+-10V, +- 2V no problem. +5-0, +-2V BIG problem.

--
W J G