Econet

[marcelaj1]

Is there any way to test if the Econet area of a BBC B is working without the server?

[WrightStuff]

ctrl+N+break will display the station number and no clock status if you have no actual network.

[BeebMaster]

You don't need a file server to use Econet, but obviously you are very limited in what you can do as you will have no shared storage, but you can still send & receive messages, remote another station, and do plenty of other things if you have locally stored utilities. For testing purposes it will be difficult to know if the Econet interface in a lone Beeb is working without a clock and another station to receive or send something.

If you can get into the Net filing system successfully (with *NET or N-BREAK) then that will at least tell you that the NFS ROM is recognised and present, and I think that the NFS ROM has detected a 68B54 controller chip, but you won't be able to go much further. Without a clock you will only get a No clock message if you try any network transmissions.

[marcelaj1]

Thanks for the info, is there a specific ROM image that I need.

[BeebMaster]

NFS 3.60 is the "latest" Acorn release NFS version for the BBC B, you can get it here:

http://mdfs.net/System/ROMs/Filing/Network/Acorn/

Have you checked if the PCB has all the Econet components fitted (top left section)?

[marcelaj1]

I took a picture of the area of the board.

What is the difference between NET and DFS,NET,I could hazzard a guess but....

[jgharston]

A simple Econet test is to just plug two machines together through a clock and do a low-level 'notify':

Code: Select all

   10 REM prog to test Econet
   20 REM SRL 01jul85
   30 
   40 INPUT "Station to send to: "stn%
   50 DIM X% 30
   60 Y%=X% DIV 256
   70 X%?0=1    :REM send string
   80 X%!1=stn% :REM Dest station
   90 $(X%+3)="*| Are you there?"
  100 A%=&14
  110 CALL &FFF1
  120 PRINT"OK"

[jgharston]

NFS 3.60 is the "latest" Acorn release NFS version for the BBC B, you can get it here:

http://mdfs.net/System/ROMs/Filing/Network/Acorn/

NFS 3.62 has a couple of bugfixes over 3.60, *EX checks for a command seperator so, eg, *EXAM matches a command 'EXAM' instead of being matched as '*EX AM'.

[BeebMaster]

I took a picture of the area of the board.
20201229_171352.jpg

What is the difference between NET and DFS,NET,I could hazzard a guess but....

All looks to be there. it may actually be a production Econet machine, as some of the chips aren't socketed, and S2 looks like it was never made. All you need is another Acorn with Econet, and a clock, and some cabling and the newest Econet will be born!

(I presume the board has been lifted up out of the case a bit, the holes in the back don't quite match.)

DFS,NET is part of the message contained in the combined Disc/Econet filing system ROM commonly known as "DNFS". Both filing systems are 8K so they were combined into a 16K ROM. It's never actually been clear to me whether the contents, DFS 1.20 and NFS 3.60, were ever officially available as separate 8K ROMs.

[marcelaj1]

My current issue 4 board is giving me problems and I figured it may be an idea for a backup. I spotted this board on ebay and it was being sold for spares and I noticed the Econet area was populated and it convinced me to take a gamble.
The board did not have any ROMs so I plugged trickys ROM in and after reseating IC6 the board fired up. I blew an OS ROM and also put in my old BASIC ROM and I had a working board, put in my old DFS ROM and I could load programs off my Gotek, the rest of the board can be tested now I have access to programs i.e. analogue, cassette, user port, serial and the Econet bit.
There is an issue with IC88 socket being faulty and I am waiting for a new socket to arrive.
I was intrigued by what I read about the"filestore" device and saw there were projects around with a view to emulating one, but until I can prove my Econet is working it is all a bit of a pipe dream.

[BeebMaster]

FileStore is a particular type of Econet file server, it's Acorn's standalone file server unit, whereas all the other Acorn file servers are an Acorn computer with Econet interface (+ additional hardware in the case of 8-bit Acorns) running file server software.

It's a lot easier to make your own file server now, as the extra hardware needed (6502 second processor and optionally ADFS-compatible hard disc) used to be significant barriers to Econet taking over the world, but now everyone can have these with PiTubeDirect and Pi1MHz SCSI.

[jgharston]

DFS,NET is part of the message contained in the combined Disc/Econet filing system ROM commonly known as "DNFS". Both filing systems are 8K so they were combined into a 16K ROM. It's never actually been clear to me whether the contents, DFS 1.20 and NFS 3.60, were ever officially available as separate 8K ROMs.

DFS 1.20 is not available as an 8K ROM as, it's bigger than 8K, it's in the high part of the ROM starting at &9F9D, it would be useless for CoPro systems as the Tube Host code is in the low part of the ROM, so any usable DFS-only ROM would be almost 9K.

NFS 3.60 is available as an 8K ROM as it's just under 7K, with the Tube code it's just under 8K, and it's just the DNFS ROM with the JSR DFS_Service NOP'd out from the service code and the string 'DFS,' removed, and a shared PrintNum routine moved to the end of the 8K.

[marcelaj1]

After a load of trouble downloading the rom, it kept opening as a page I managed to get the image by switching from Firefox to another browser.
Loaded into boobip device and did a Ctl-N-Break and got:

"Econet Station 050 No Clock"

Is this just the ROM talking or does my beeb have a "working" Econet with nothing to talk to?

[jgharston]

After a load of trouble downloading the rom, it kept opening as a page I managed to get the image by switching from Firefox to another browser.
Loaded into boobip device and did a Ctl-N-Break and got:

"Econet Station 050 No Clock"

Is this just the ROM talking or does my beeb have a "working" Econet with nothing to talk to?

It's stating that there's no clock. That could be because you have no clock, or you have a clock but you haven't plugged it in, or you have a plugged-in clock but it's broken, or you have a working plugged-in clock, but the network cable is broken, or you have a working, plugged-in clock with working cables, but the network clock receivers aren't working.

[jgharston]

After a load of trouble downloading the rom, it kept opening as a page

Doesn't the browser you use have a 'save target file' option? Every browser I've ever used over the past 25 years have.
Eg:

[marcelaj1]

Doesn't the browser you use have a 'save target file' option? Every browser I've ever used over the past 25 years have.

It used to come up on a right click, but does not any more, I remember downloading other stuff from the site with no problems, but since the last update the browser is getting anoying - I will have to spend some time reading up on Firefox settings.

It's stating that there's no clock. That could be because you have no clock, or you have a clock but you haven't plugged it in, or you have a plugged-in clock but it's broken, or you have a working plugged-in clock, but the network cable is broken, or you have a working, plugged-in clock with working cables, but the network clock receivers aren't working.

I don't have anything plugged in the Econet socket, so I was expecting the "no clock" bit, I was wondering if the Econet bit on the board was OK, would the system report back a station number if it was broken.

[WrightStuff]

I don't have anything plugged in the Econet socket, so I was expecting the "no clock" bit, I was wondering if the Econet bit on the board was OK, would the system report back a station number if it was broken.

Beebmaster answered this earlier :

If you can get into the Net filing system successfully (with *NET or N-BREAK) then that will at least tell you that the NFS ROM is recognised and present, and I think that the NFS ROM has detected a 68B54 controller chip, but you won't be able to go much further. Without a clock you will only get a No clock message if you try any network transmissions.

So basically you can say you have a working NFS ROM and a 68B54 has been detected.
Doesn't mean necessarily it will ok when you plug it into an actual network, but this is probably about the limit of what you can validate at this point without one.

[marcelaj1]

This is my dilemma, I don't want to throw money at this if the Econet area of the board is broken. A proper clock will set me back £50, so the next option is to bodge one.
Reading some of the docs on the net am I correct in thinking...
I would need to create a mark space waveform with the space being about 3 times longer (roughly just to test it out) than the mark and send that out on the clock +ve leg and an inverted version on the clock -ve leg. That being the case what would be the minimum frequency of the clock, I read it can be as low as 17Khz.
I need to terminate the data lines - a set of passive components that I probably have in my box of bits, although as this network would be across my desk not sure I really need it/them.

[marcelaj1]

Threw an oscilator together using a 555 and set it to 20/80 mark space at 120Khz, pushed it through a SN75176BP to a passive terminator, all built on breadboard. Plugged it into the econet port and did a CTL-N BREAK, I got the message that the station number was 50, ah I thought the "no clock" message has gone so it may be happy....
Tried "*I AM ALIVE" and after a few seconds it came back with "line jammed" - from what I have been reading this can be a network fault (unlikely as all there is on it, is a passive terminator) or something in the econet bit of the beeb.

Got the scope out and looked at IC93 pin 11, permanently high, so I pulled the chip and looked at IC89 pin 6, permanently high - does this indicate a faulty IC89?

[IanS]

No, I suspect your terminator is wired up wrong.
It will only attempt to transmit if it sees the network is quiet. After a timeout is has decided the line is premanently busy (jammed).
What circuit ar you using for the terminator. Got any pictures of your setup.

[marcelaj1]

The terminator is based on this

I put the resistors and capacitor on the breadboard.

Are there any valid reasons for pin 6 of IC89 (TXD) to be permanently high if I pull IC93 as that would mean that 89/6 is open cct nothing connected, so the high is coming from IC89.

The oscilator is based on this cct

I got the mark/space and frequency where I wanted and measured the values and used resistors rather than variable ones.

Here is a picture of the spaghetti

The clock terminator is on the big board just after the SN75176BP on the left and the data terminator is on the little board on the left. The white cable goes to th beeb.

[IanS]

Do you have the pin numbering correct, 5-pin DIN's are strange.
What do you measure on pin 3 (RxD) of the ADLC (IC89)?

[BeebMaster]

Screenshot_20210111-184649_Chrome.jpg

product-placement-p-logo.jpg (4.36 KiB) Viewed 713 times

!!!

[marcelaj1]

Sorry should have re-named the file and cropped the image

[IanS]

The terminator is based on this
20210115_190639.jpg
I put the resistors and capacitor on the breadboard.

...

and the data terminator is on the little board on the left. The white cable goes to th beeb.

The terminator board appears to only have 3 wires going to it. To correctly bias the data lines it needs to pick up a supply from the clock lines.

[marcelaj1]

It looks like one of the jumpers is missing, I will put it back and see what the readings are.

[marcelaj1]

I remembered I took the jumper off and added a capacitor to isolate the clock from the data just in case there was interference, put it back to standard now.
Took measurements with my scope (ICno./pin)
89/2 - RTS -3.6v solid
89/3 - RXD - 0.3v solid
89/4,5 - RXC,TXC - 4.7v pulses in time with CLK+
89/27 - DCD - 0.2v solid
89/28 - CTS - 4.2v pulses not always in line with CLK+

87/1 - INPUT A - 4,7v pulses in time with CLK+
87/4 - !Q - 0.2v solid
87/13 - Q - 4.2v solid
87/14 - 0.1v solid
87/15 - 0.7-1.2v sawtooth waveform

The cct diagram (bbc_b_schematic_issue_4_hi_res.png) is not clear with IC87 pins 3 & 2, it shows them joined together and intimates they are tied to +5 (as opposed to R50 for example, which is specific), I assume this is correct?

Looking at the data sheet for IC87, if pins 2(INPUT B) and 3(CLR) are both high then I should get a LHL on Q (pin 13) and a HLH on !Q (pin 4) in time with the CLK on pin 1(INPUT A) is this correct? I get a sawtooth waveform on pin 15, so I assume some sort of trggering is going on, could I have a broken IC87?

[IanS]

89/3 - RXD - 0.3v solid

Did you check your terminator wiring. On an idle network, RXD should be high. So currently it thinks the line is jammed.

edit:
viewtopic.php?p=161029#p161029
viewtopic.php?p=161038#p161038

[marcelaj1]

Did you check your terminator wiring. On an idle network, RXD should be high. So currently it thinks the line is jammed.

I quickly took out the terminator on the breadboard and put in the biasing resistors you pointed out, and issued a "*I AM ALIVE" and after about 15 seconds got a "Net error", would this be because nothing answered?

[IanS]

Did you check your terminator wiring. On an idle network, RXD should be high. So currently it thinks the line is jammed.

I quickly took out the terminator on the breadboard and put in the biasing resistors you pointed out, and issued a "*I AM ALIVE" and after about 15 seconds got a "Net error", would this be because nothing answered?

With NFS 3.34 I get "Not Listening" with just a clock (and termintor). Not sure if different NFS versions give different error messages or you still have other problems.