Fixed: Mode 7 intermittent lines

for bbc micro/electron hardware, peripherals & programming issues (NOT emulators!)
roganjosh
Posts: 119
Joined: Sat Dec 10, 2016 6:51 pm
Location: W.Yorks
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by roganjosh » Sun Aug 25, 2019 5:14 am

Good stuff. Well done.
Also, as you're very probably aware, a little wobbling of the display on changing mode is normal behaviour for many HDMI converters.
I go for the low profile crystals as they are unlikely to be NOS ones.

Alan
Last edited by roganjosh on Sun Aug 25, 2019 6:55 am, edited 1 time in total.

User avatar
retroclinic
Posts: 3032
Joined: Thu Jul 03, 2008 1:22 pm
Location: East Riding of Yorkshire
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by retroclinic » Wed Sep 11, 2019 9:13 am

Interesting read, as I've also had jumping from several beebs using these ElCheapo converters. The 16MHz crystal you used, I looked at the eBay ad, but it doesn't mention the tolerance. Closest I can find in an HC49 case is 30ppm, I'm presuming this will be a more accurate version than that originally supplied 35+ years ago?

A note on the background lines seen on one of the photos when used with a SCART TV. I've seen this often, and it's a grounding issue. Most modern LCDs do not connect to earth, being either a 2 pin mains input, or via an AC-DC adapter with a 2 pin input. As the signal is TTL from the Beeb, reducing the brightness and increasing the contrast gets rid of this.
Image

User avatar
KenLowe
Posts: 799
Joined: Mon Oct 18, 2004 4:35 pm
Location: Scotland
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by KenLowe » Thu Sep 12, 2019 11:44 am

retroclinic wrote:
Wed Sep 11, 2019 9:13 am
Closest I can find in an HC49 case is 30ppm
There's one seller quoting 20ppm:

https://www.ebay.co.uk/itm/IQD-16MHz-Cr ... Sw0JFbVeoM

User avatar
retroclinic
Posts: 3032
Joined: Thu Jul 03, 2008 1:22 pm
Location: East Riding of Yorkshire
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by retroclinic » Thu Sep 12, 2019 1:13 pm

Ok, been doing some testing on this today. I've tried several different new 16MHz crystals from different manufacturers in place of existing original crystals that were causing the re-sync issue on the converter, and any of the new type I've tried have been fine, and they've all been 30ppm. I've had a screen next to me while watching some TV with a test program running, and not seen a single flash up of a resync from the converter.

Thanks for the eBay link for those crystals, but at 10x the price of even RS (who are usually in themselves expensive), I'll pass on those, and I really don't trust they're either new or the stated tolerance.

Interesting looking into this, and reading up on how crystals degrade over time. I'd never considered replacing the main clock crystal as a matter of course on all machines, as I've always tested them on a CUB, which as with any CRT is very forgiving of erroneous clock speeds. However, as everything original in Beebs are now 30-35 years old, I will replace them from now on for the sake of 30p.

They say you learn a new thing every day!

Thanks, Mark.
Image

User avatar
anightin
Posts: 368
Joined: Thu Aug 23, 2018 1:03 pm
Location: Cambridge UK
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by anightin » Thu Sep 12, 2019 2:22 pm

retroclinic wrote:
Thu Sep 12, 2019 1:13 pm
Ok, been doing some testing on this today. I've tried several different new 16MHz crystals from different manufacturers in place of existing original crystals that were causing the re-sync issue on the converter, and any of the new type I've tried have been fine, and they've all been 30ppm. I've had a screen next to me while watching some TV with a test program running, and not seen a single flash up of a resync from the converter.

Thanks for the eBay link for those crystals, but at 10x the price of even RS (who are usually in themselves expensive), I'll pass on those, and I really don't trust they're either new or the stated tolerance.

Interesting looking into this, and reading up on how crystals degrade over time. I'd never considered replacing the main clock crystal as a matter of course on all machines, as I've always tested them on a CUB, which as with any CRT is very forgiving of erroneous clock speeds. However, as everything original in Beebs are now 30-35 years old, I will replace them from now on for the sake of 30p.

They say you learn a new thing every day!

Thanks, Mark.
You know what this means. I guess were now obliged invent a term now on the forum, and indeed maybe within the wider retro community along the lines of re-capping.

Retro-Reclocked ? (coz 'Recrystallization' is far too fancy a word IMHO)

:D
Last edited by anightin on Thu Sep 12, 2019 2:28 pm, edited 3 times in total.

Kazzie
Posts: 1025
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by Kazzie » Thu Sep 12, 2019 2:58 pm

anightin wrote:
Thu Sep 12, 2019 2:22 pm
You know what this means. I guess were now obliged invent a term now on the forum, and indeed maybe within the wider retro community along the lines of re-capping.

Retro-Reclocked ? (coz 'Recrystallization' is far too fancy a word IMHO)

:D
Requartzing? :)
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM, ZIDEFS with 512MB CF card
Acorn System 1 home-made replica

User avatar
KenLowe
Posts: 799
Joined: Mon Oct 18, 2004 4:35 pm
Location: Scotland
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by KenLowe » Thu Sep 12, 2019 3:13 pm

retroclinic wrote:
Thu Sep 12, 2019 1:13 pm
However, as everything original in Beebs are now 30-35 years old, I will replace them from now on for the sake of 30p.
Out of interest what crystal will you be using? Have you tried any of the HC-49US (I assume the S stands for small) package devices in place of the original HC-49U? The US package seems to be cheaper and more readily available.

User avatar
retroclinic
Posts: 3032
Joined: Thu Jul 03, 2008 1:22 pm
Location: East Riding of Yorkshire
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by retroclinic » Thu Sep 12, 2019 3:21 pm

KenLowe wrote:
Thu Sep 12, 2019 3:13 pm
retroclinic wrote:
Thu Sep 12, 2019 1:13 pm
However, as everything original in Beebs are now 30-35 years old, I will replace them from now on for the sake of 30p.
Out of interest what crystal will you be using? Have you tried any of the HC-49US (I assume the S stands for small) package devices in place of the original HC-49U? The US package seems to be cheaper and more readily available.
I've tried both types, they work the same. RS do the low profile ones that are 20p each (in a bag of 100), Farnell do full height Multicomp ones that are a bit more, but yoiu can buy just ones. Both work just fine.

Thanks, Mark.
Image

User avatar
KenLowe
Posts: 799
Joined: Mon Oct 18, 2004 4:35 pm
Location: Scotland
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by KenLowe » Thu Sep 12, 2019 3:39 pm

Thanks for the info. I'm guessing you tried the RS 146-3926. Given that I don't need that quantity, I've just ordered up a pack of 10 RS 814-9440 for £1.70 (including VAT & postage). Ony real difference I can see between the two is load capacitance, and frequency tolerance / stability. Assuming they arrive tomorrow, I'll test and provide an update tomorrow.

User avatar
retroclinic
Posts: 3032
Joined: Thu Jul 03, 2008 1:22 pm
Location: East Riding of Yorkshire
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by retroclinic » Thu Sep 12, 2019 3:50 pm

KenLowe wrote:
Thu Sep 12, 2019 3:39 pm
Thanks for the info. I'm guessing you tried the RS 146-3926. Given that I don't need that quantity, I've just ordered up a pack of 10 RS 814-9440 for £1.70 (including VAT & postage). Ony real difference I can see between the two is load capacitance, and frequency tolerance / stability. Assuming they arrive tomorrow, I'll test and provide an update tomorrow.
Yes, I had, those others are even cheaper, yes just watch the load capacitance. I'm not sure it'll make a lot of of difference, just the original caps in the beeb are 30pf, so they may need changing out for 18pf, or may be fine as it is.

Let us know how you get on.

Thanks, Mark.
Image

User avatar
KenLowe
Posts: 799
Joined: Mon Oct 18, 2004 4:35 pm
Location: Scotland
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by KenLowe » Fri Sep 13, 2019 1:27 pm

retroclinic wrote:
Thu Sep 12, 2019 3:50 pm
KenLowe wrote:
Thu Sep 12, 2019 3:39 pm
Thanks for the info. I'm guessing you tried the RS 146-3926. Given that I don't need that quantity, I've just ordered up a pack of 10 RS 814-9440 for £1.70 (including VAT & postage). Ony real difference I can see between the two is load capacitance, and frequency tolerance / stability. Assuming they arrive tomorrow, I'll test and provide an update tomorrow.
Yes, I had, those others are even cheaper, yes just watch the load capacitance. I'm not sure it'll make a lot of of difference, just the original caps in the beeb are 30pf, so they may need changing out for 18pf, or may be fine as it is.

Let us know how you get on.

Thanks, Mark.
That's me received the RS 814-9440 crystals, and I've installed one into one of my beebs that would cause the occasional glitch with the HDMI adaptor. It's been running now for over 30mins without a single glitch. Seems to have solved it for me. Similar to you Mark, any time I've got a board out of the case, I'll replace the crystal as a matter of course.

User avatar
maniacminer
Posts: 33
Joined: Thu Sep 21, 2017 1:59 am
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by maniacminer » Wed Oct 02, 2019 12:24 am

Interesting topic. I had a VideoNULA in my Beeb and it shimmered like crazy in MODE7. It would start off ok, then gradually over a few minutes start losing pixel clock. I never thought to change the 16MHz crystal :o I am wondering if this is like a RIFA scenario, replace on sight? I'll have to give the VideoNULA another try! I am mystified by R102 and C42 on the input of IC40, one input of the two input NAND gets the 16MHz clock and the other input is connected to the 16MHz clock and this little RC circuit. Anyone know why this is fitted? Is it a Furber finger of god mod?

roganjosh
Posts: 119
Joined: Sat Dec 10, 2016 6:51 pm
Location: W.Yorks
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by roganjosh » Wed Oct 02, 2019 6:15 am

maniacminer wrote:
Wed Oct 02, 2019 12:24 am
I am wondering if this is like a RIFA scenario, replace on sight? I'll have to give the VideoNULA another try!
When I first pointed out this fix in a different thread I noted that the problem has affected all my Model Bs but, so far, none of my Master 128s. The oscillator has the same linear design (IC43 74S04 gates) in both machines, however there are resistor & capacitor differences. Perhaps that contributes. As I said, I treat this as a stock fault on Bs and replace on sight. I've held off doing so on Masters so far, though that's only a sample size of 2. However, I've a loose Master mobo to check plus another Master 128 arriving later this week so I can see whether the distinction holds true. I'd be interested to hear whether anyone with a Master has experienced the problem.

Alan

User avatar
retroclinic
Posts: 3032
Joined: Thu Jul 03, 2008 1:22 pm
Location: East Riding of Yorkshire
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by retroclinic » Wed Oct 02, 2019 6:46 am

The Master has a different SYNC output, it doesn’t blank the h during the v period, it inverts it, so it could be that difference making the Master more stable?

Disclaimer - this is from memory as I’m waking up, so I reserve the right to correct myself! :lol: :lol:
Image

roganjosh
Posts: 119
Joined: Sat Dec 10, 2016 6:51 pm
Location: W.Yorks
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by roganjosh » Thu Oct 03, 2019 3:52 pm

maniacminer wrote:
Wed Oct 02, 2019 12:24 am
I am mystified by R102 and C42 on the input of IC40, one input of the two input NAND gets the 16MHz clock and the other input is connected to the 16MHz clock and this little RC circuit. Anyone know why this is fitted? Is it a Furber finger of god mod?
The intent of the circuit appears to be to improve the clock waveform by shortening the rise and fall times. Owing to the time taken to charge the capacitor the nand gate output will toggle at a higher input voltage than normal. The converse happens on the other edge as it takes a little time for the capacitor to discharge.

User avatar
maniacminer
Posts: 33
Joined: Thu Sep 21, 2017 1:59 am
Contact:

Re: Fixed: Mode 7 intermittent lines

Post by maniacminer » Thu Oct 03, 2019 7:24 pm

roganjosh wrote:
Thu Oct 03, 2019 3:52 pm
maniacminer wrote:
Wed Oct 02, 2019 12:24 am
I am mystified by R102 and C42 on the input of IC40, one input of the two input NAND gets the 16MHz clock and the other input is connected to the 16MHz clock and this little RC circuit. Anyone know why this is fitted? Is it a Furber finger of god mod?
The intent of the circuit appears to be to improve the clock waveform by shortening the rise and fall times. Owing to the time taken to charge the capacitor the nand gate output will toggle at a higher input voltage than normal. The converse happens on the other edge as it takes a little time for the capacitor to discharge.
Thanks for the explanation! I must take a look at these two clock lines with the 'scope and see if they are stable and actually 16MHz...

Post Reply