• Please review our updated Terms and Rules here

Newbrain

Gary C

Gary C

Veteran Member
Joined
May 26, 2018
Messages
2,399
Location
Lancashire, UK
Next on the bench is a Grundy Newbrain

Circuit diagrams are here http://www.goffart.co.uk/museum/schematics/newbrain.pdf

PSU was powered up by one of the volunteers to a Big Bang and lots of RIFA smoke :)

The PSU is a simple transformer/rectifier with three transistors roughly regulating using Zeners as a voltage reference. RIFA removed (a chunky affair with three terminals as it actually contains three capacitors, between L&N and two to earth)
Unit powered up again and all three supplies present.

Opening up is a pain, lots and lots of screws to reveal three circuit boards. I can see why this was overpriced compared to say the Spectrum. Lots and lots of discrete logic, so much that the RAM can't fit on the motherboard and has to be farmed off to a daughterboard.

Plugged into the NB, and... No 5v

The regulation uses an OPAMP and a precision voltage source, using the amp as a basic proportional controller. +5v rises above the reference which tends to lower the output of the op amp which reduces the bias on the regulator transistor which reduces the +5V.
Voltage on the reference should be 2.45V but is 1.2V ?
This machine has a slight modification over the drawing and has a the 8V DC then a 100ohm resistor then a 1.2k resistor then the reference, oddly the 100ohm resistor was dropping about 7v across it and the 1.2k resistor only about 0.2v which means they cannot have the same current flowing through them. Simple fault, the capacitor to 0V has failed with a low resistance to 0v so much so that the reference isn't getting enough to operate. Cap replaced and gunk wiped of and voltage restored.

I'm not a fan of 'Recapping' but looking at the electrolytic on this board, they are all leaking, so next job is replacement.
 
Yep !

The classroom stinks and so does the spare bedroom :(

Slowly dissipating :)

All but two Electrolytics replaced (refresh of stocks on order) and keyboard reconnected. Display on video is rock solid. Probably need to find or create some diagnostic/memory tests next.
Edit, except that the new brain seems to have quite a good memory check built into the ROM and its showing 32K of working memory.

but other than that, its a working machine. The PCB's look very similar quality to my Samsung made spectrum (Issue 4S).
 
Last edited:
.. wasn't there some software with it? I think it was all on Cassette though. Did any one ever make floppies for the Newbrain?
.. looking at the Newbrain history I'm amazed any ever got made...
 
Cassettes are at the museum so can't get to them until the weekend but I am surprised that I can't find any WAV files or the like online. There is an emulator but seems to store the 'tape' files as text or bin files
 
Interesting

The newbrain has two RC type 'reset' lines, one is for the CPU and one is called PWRUP


1725379486752.jpeg

The bottom one resets the CPU, the top one is the PWRUP circuit and acts (obviously) later and in the opposite sense.

However, while the junction of resistor 128 & capacitor 125 at power on, start at 5V, then fall to ~0V, the junction of 129 and 127 only falls to about 1.7V and its not always enough to trigger the 74C14 inverting Schmitt gate. Replaced the Capacitor to no avail.
I can seriously reduce the resistor value to force it to fall below the operating point, but I can't quite work out why this would happen and don't really want to leave it with a much lower resistor in place (talking about 10K to make it work reliably) as the current is coming from somewhere and I will only probably have to open its (really daft) case again when it finally fails.

Anyone seen similar ?

I have a new IC on order in case its somehow holding itself up (which is about the only thing I can think it is) but it does seem odd.
 
Ah !

Took the IC out and it still did the same. No other source of power yet the junction between 129 and 127 sits at 1.7V. Lifted that junction from the PCB and it falls to 0V so there is something tracking on the board. Even a good clean hasn't resolved it suggesting that their is some damage from the leaking capacitor. It is only 3uA but it was enough to stop it working.

Bit the bullet and dropped R129 to 390K and it now works. Will have to see if it degrades further over time.
 
Last edited:
Gary C said:
Ah !

Took the IC out and it still did the same. No other source of power yet the junction between 129 and 127 sits at 1.7V. Lifted that junction from the PCB and it falls to 0V so there is something tracking on the board. Even a good clean hasn't resolved it suggesting that their is some damage from the leaking capacitor. It is only 3uA but it was enough to stop it working.

Bit the bullet and dropped R129 to 390K and it now works. Will have to see if it degrades further over time.
Click to expand...
Capacitor damage is a new thing to me. Have a MicroVax PSU that died through it. Looks like we have to remove batteries and caps now before storing...
 
Certainly have seen track damage from capacitors in a Mac LC and have seen discolouration in monitors but not enough to stop circuits working after cleaning.
Working it out, I have 1 Mohm path from +5v through the PCB. Maybe it would get better if I really went to town on cleaning but 560k is probably too large for this circuit.

Really, if I hadn't already put the case together, I might have considered dropping the value of R128 as well as R129 to make a snappier reset.
 
So, dug the old TEAC hifi tape drive out and trying to load tapes into the PC so I can reliably load them later.

Got the first one recorded as a mp4 successfully but its a right pain coping with different levels and cleaning dirty heads but I can't find anywhere to download them from.

So will just have to do them step by step.
 
Well this is odd

Copied three programs from tape into wav files.

The first (short) program loads fine, but the second two, not so much.

The files are separated into blocks with a tone leader followed by data, then silence then the next block. As the data is loading, the screen flickers which is a good indication its reading the tape. For the first three blocks everything is fine, but starting block four, after the leader, no flicker. Now the odd thing, if I step back to the silent portion and let it run, the block now loads, then fails on the next block, rewind that and it too loads. Do this on all remaining blocks and the program loads fine.

With the largest program in memory, saving this directly from the newbrain to the PC as a WAV then load it back and it does the same thing even though the resulting waveform is much clearer and with constant amplitude.

I increased the silent period and the leader and that at first seemed to help, but not on the bigger file. Its the stopping of playback that seems to make the difference.

Can't quite work it out but I wonder if the PC is doing something to the line (DC bias ?) when playing that's dropped when stop or pause is clicked. Going to have to open it up to scope it, but first I think I will try to resurrect the other unit and see if that does the same.
 
I have found unbalanced PC to anything else connections to be very noisy - possibly due to earth loops.

Designing the audio and visual system at our Church, most of our problems are down to earth loops of one form or another. I found the other month that all of our SDI video problems were due to Class 2 TVs and USB power sockets. In fact, I got a mild electric shock off the screen of a disconnected SDI cable (at one end) where I measured 40 V AC on the screen relative to real earth. A main plug, some green and yellow striped earth cable and an SDI connector had everything well and truly strapped down to real earth, and all of our video problems have gone away!

We used to get similar problems with unbalanced audio equipment. PCs and guitarists are the worst - where I use transformer passive isolation and in-built ground lift switches to cure the problems at source.

I will be up at your place on Sunday evening ready for an important meeting on Monday morning...

Dave
 
The odd thing is the way it works perfectly if I pause the audio during the quite or sync tone leader, its also odd that it work fine every time for the initial and second blocks of data so its not noise as such corrupting the data.

Sparked up a second newbrain, and it does exactly the same when loading.
 
Numpty time again.

The Newbrain isn't that fast and when its loaded in a block from tape, it takes time to do some processing. While it does this it opens the tape control line to stop the tape until its ready to read it again.

Simple stuff but using a PC without a motor control its easy to forget :)
 
And for info if anyone else stumbles across this

The Newbrain cassette lead only puts a single wire to the ear lead, that means you have to plug the mic lead in too so that you get a ground connection.

Thats fine, except when trying to use a tapuino type device that only has the mic output and therefore no connection to ground. Custom lead required.
 
Really not happy with the heat the 5V regulation circuit puts out. The regulator transistor gets hot and the PSU brick also gets almost to hot to touch.

1727542161552.png
So, out with resistors 27 and 14 plus transistor 22. TRACO 1-2450 Vin connected to Emitter connection of 22 and Vout connected to Collector of 22 then Gnd connected to any convenient ground.

Powered up and +5 now sits at 4.95V stable and cool. The PSU brick (which uses Zener's as a reference to bias some TIP32's) also runs nice and cool.

Its a bit of a lashup for now (shame the pinouts of the regulator don't align with the vacant via's of Transistor 22 but can't have everything. Probably should remove Transistor 17 too
 
Last edited:
You must log in or register to reply here.
Back
Top