exxos's DFB1 trials
-
TotO
- Posts: 36
- Joined: 04 Mar 2021 17:27
Re: exxos's DFB1 trials
I think that is not very complicated to handle multi push/shove while dragging only one IC at the time. May be for a next version... I will like first to be able to route a full bus signals, as the schematic handle this kind of element since decades on Eagle and Kicad. :D
-
exxos
- Site Admin

- Posts: 28391
- Joined: 16 Aug 2017 23:19
- Location: UK
Re: exxos's DFB1 trials
@Badwolf Room to fit a 40MHz OSC on there as well. I guess a Plan B: would be good if things still go wonky with 50MHz :shrug:
You do not have the required permissions to view the files attached to this post.
-
exxos
- Site Admin

- Posts: 28391
- Joined: 16 Aug 2017 23:19
- Location: UK
Re: exxos's DFB1 trials
I think I will fit both oscillators. Because people may use their own CPU with the thing which may not even be 50MHz capable to start with. Plus with all the changes I am doing, it does not guarantee success either. Though at this point we still don't know if the FPU's themselves can even run at 40MHz or higher speeds reliably. At least I will give it the best chance possible into reaching those higher speeds. That is all I can realistically do.
-
exxos
- Site Admin

- Posts: 28391
- Joined: 16 Aug 2017 23:19
- Location: UK
Re: exxos's DFB1 trials
@Badwolf The LCSC part numbers you have, are they part numbers directly from LCSC site, or the LCSC part numbers from JLCPCB ?
I ask because sometimes the numbers do not always tally up or get discontinued on one site but not the other etc
I ask because sometimes the numbers do not always tally up or get discontinued on one site but not the other etc
-
exxos
- Site Admin

- Posts: 28391
- Joined: 16 Aug 2017 23:19
- Location: UK
Re: exxos's DFB1 trials
Currently looking for oscillator options in LCSC.. One thing I did not the realise is there are two variants..Symmetry (duty cycle)
Basically a 40-60 ratio, or 45-55 ratio. I guess this could explain why some of the oscillators seem really " out of spec".
Ideally they should be 50:50 ratio, But so far in my searching such a thing does not simply exist.
I guess duty cycle can also be very dependent on what is seen as a logic high and logic low as well :roll:
Problem is the datasheet does not specify a part number variation for the different duty cycles :roll:
Basically a 40-60 ratio, or 45-55 ratio. I guess this could explain why some of the oscillators seem really " out of spec".
Ideally they should be 50:50 ratio, But so far in my searching such a thing does not simply exist.
I guess duty cycle can also be very dependent on what is seen as a logic high and logic low as well :roll:
Problem is the datasheet does not specify a part number variation for the different duty cycles :roll:
You do not have the required permissions to view the files attached to this post.
-
exxos
- Site Admin

- Posts: 28391
- Joined: 16 Aug 2017 23:19
- Location: UK
Re: exxos's DFB1 trials
So it looks like the JWT brand is the only one which specifies the duty cycle as 45-55%
So the idea is have two oscillators which are selectable via a inverter and a jumper link. So this way should be quickly changeable between the two oscillators.
I have not quite worked out how to fit the jumper link in yet... Geez those things are huge! :lol: :roll:
So the idea is have two oscillators which are selectable via a inverter and a jumper link. So this way should be quickly changeable between the two oscillators.
I have not quite worked out how to fit the jumper link in yet... Geez those things are huge! :lol: :roll:
You do not have the required permissions to view the files attached to this post.
-
exxos
- Site Admin

- Posts: 28391
- Joined: 16 Aug 2017 23:19
- Location: UK
Re: exxos's DFB1 trials
Does anyone have any ideas about why I cannot move some of those blue tracks ? I'm really baffled because the red track to the left I can move.. But the via and the blue track itself just refuses to move. I have not had any other problems moving trucks elsewhere on the board. So have no idea what is going on :(
EDIT:
I figured it out. The tracks on the connector on the far right are not actually going to the correct pads. This is because I moved it at some point in the tracks did not follow the pads (which is seriously irritating") so the pads are actually shifted by one compared to what the tracks are supposed to be connected to :roll:
EDIT:
I figured it out. The tracks on the connector on the far right are not actually going to the correct pads. This is because I moved it at some point in the tracks did not follow the pads (which is seriously irritating") so the pads are actually shifted by one compared to what the tracks are supposed to be connected to :roll:
You do not have the required permissions to view the files attached to this post.
-
exxos
- Site Admin

- Posts: 28391
- Joined: 16 Aug 2017 23:19
- Location: UK
Re: exxos's DFB1 trials
I have somehow ended up in a situation with the power rails :(
In the schematic there is VCC and VCC3. But for some reason on the PCB they are both connected together ?! Anyone know why this could happen ?!
It is also strange that VCC3 is missing from the net list :shrug:
:shrug:
Even if I deleted the whole regulator section, the first error vanishes, but the PCB is still not showing the correct VCC3 nets :(
EDIT:
So the problem seems to be with my changes relating to the oscillators. If I delete the whole section then the PCB them behaves properly :stars:
Seems to be the inverter block VCC3 :shrug:
In the schematic there is VCC and VCC3. But for some reason on the PCB they are both connected together ?! Anyone know why this could happen ?!
It is also strange that VCC3 is missing from the net list :shrug:
:shrug:
Even if I deleted the whole regulator section, the first error vanishes, but the PCB is still not showing the correct VCC3 nets :(
EDIT:
So the problem seems to be with my changes relating to the oscillators. If I delete the whole section then the PCB them behaves properly :stars:
Seems to be the inverter block VCC3 :shrug:
You do not have the required permissions to view the files attached to this post.
-
exxos
- Site Admin

- Posts: 28391
- Joined: 16 Aug 2017 23:19
- Location: UK
Re: exxos's DFB1 trials
I think I have sort of figured it out...
The actual part is a "up to" 5V part. I think it just must have been inherently connecting it to VCC by itself.
I swapped it for a proper LVC (3.3v) part and now the problem has gone away it seems :shrug:
Not really sure if that is a "feature" or a bug. Though as it has power rails which I can put my own label on, I would say that some odd bug in KiCad :roll:
The actual part is a "up to" 5V part. I think it just must have been inherently connecting it to VCC by itself.
I swapped it for a proper LVC (3.3v) part and now the problem has gone away it seems :shrug:
Not really sure if that is a "feature" or a bug. Though as it has power rails which I can put my own label on, I would say that some odd bug in KiCad :roll:
-
exxos
- Site Admin

- Posts: 28391
- Joined: 16 Aug 2017 23:19
- Location: UK
Re: exxos's DFB1 trials
The next change is to move the regulator the centre of all the 3.3V chips.
I will use a chunk of the bottom layer as a heatsink for the regulator as well. Not really sure how hot the thing gets but it's "free copper" anyway.
Not sure why I haven't got a 3-D FPU socket :shrug:
I will use a chunk of the bottom layer as a heatsink for the regulator as well. Not really sure how hot the thing gets but it's "free copper" anyway.
Not sure why I haven't got a 3-D FPU socket :shrug:
You do not have the required permissions to view the files attached to this post.
Who is online
Users browsing this forum: ClaudeBot and 1 guest