Heathkit IG-18 update and improvement

[pizzigri]

Hello everyone,

So I managed to get a fairly clean Heathkit IG-18, a very early model according to the board revisions.  So, I just looked round EEVBlog to find some info on how to improve it, and in the end found the excellent, but ageing, Tronola website, with the Super IG18SL design, by Steve Lafferty Modifications: Heathkit IG-18 Audio Generator
and
http://www.tronola.com/IG-18SL_Construction_Notes.pdf
Larry Burk designed a PCB in Eagle and the files are downloadable form Tronola website.
Dick Moore experimented with the design and also his results are hosted in this website.
However, the info is really dated, many components are completely obsolete and unobtanium - not to say that the specific rotary switch in the BOM is crazy expensive.
I read through all the various updates, including Dick Moore's experiments, and have come to the conclusion that a basic update of the board could be attempted. So I downoladed DipTrace... mind you, I am totally green in using it.
I tried to update the power supply to something actually modern.
The PS built in Lafferty's board involves powering the device with a tranformer wired for 240V but connected to 110, which is very unpractical for me, as I am in EU (Italy). Cant remember where I pilfered it, but I just needed a schematic and chose the LM317T and LM337T (both high voltage versions so they accept the rather high voltage coming from the trafo).
I also rerouted all top traces so that they should work as an actual layer and not as a simple guide to add top patch wires.
Here's the deal: I have absolutely no idea if I did a good job. I tried looking at most rules for grounding loops and other obvious errors, but am not competent enough to see if i made mistakes or not. I manually corrected lots of autoroute's choices, redrawing many times the placement of components in the schematic inside DipTrace to improve it, however I need your help to check it out.
So, I'm posting the gerber and diptrace file... it is a work in progress and I really do not advise to use it. Also, the schematic for the PS, while the schematic for the original Lafferty's design can be downloaded from Tronola.

In reality, my objective is to redesign completely the circuit to use SMD components, possibly 1206, as I have seen that they are easy to solder by hand (and even easier by using even a crappy 858D). In this case, my dream would be to have the PS in a separate board, sitting where the original was in the IG-18, and use the real estate on the OSC board to place all the range caps and AGC caps needed, that in both Lafferty and Moore projects are fashioned around the switch. Problem is, I am no EE just a hobbyst and have nowhere near the capability to do so.
Can anyone help me?

[ledtester]

When using the LM317/337 be aware of the minimum load requirement for regulation (it's like 10mA).

[pizzigri]

Duly noted...
Can you spot obvious mistakes in the board? Any suggestions on how to improve it?

[ledtester]

I don't have DipTrace installed, but I did notice that your Ground point only has 5 traces coming out from it where are the original schematic and wiring guide has 7.

It probably won't make a big difference, but perhaps you can thicken up those ground traces a bit. You can also go directly from C24 to C22. Not important but it'll look nicer.

[pizzigri]

OK, doing that. I'm also updating the PCB to reduce components around U1, since the Harris Opamp is not used anymore. Would you have a source for the
C4D0604N-A 6P4T rotary switch? Cant find it at a resonable price (or with 30$ of shipping)...

[ledtester]

Those C4D0604N-A switches are expensive! Even in the US they're $40! It's tempting to consider some sort of electronic switching circuit.

I suppose there's always aliexpress...

"1PCS 3 Deck 6mm Shaft Band Channel Rotary Switch Selecter 3P6T/3P7T/3P8T/3P9T/3P10T/3P11T/6P3T/6P4T/6P5T/9P3T/12P2T Ceramic"
https://www.aliexpress.com/item/1005001664740132.html

Not as pretty, but a lot cheaper. Make sure the dimensions will work for you.

[pizzigri]

I updated the PCB and refined it a bit. Also, I managed to export it to Eagle so you don't need DipTrace to open the file.
It does not give out errors any more (well apart from some due to manual routing), can someone check it out?
If someone can validate it, I'd be happy.