after looking around for the many other topics, I finished what to think and try for my problem, so I'm now humbly asking the opinion of other fellow astrophotographers to see if I can figure out what's wrong with my (brand new!) EQ6R-pro.
I have a permanent setup outside, on a Skywatcher pier tripod. I used this setup for years with an HEQ5 and it worked brilliantly, despite the HEQ5 was sometimes guiding more than 14kg of stuff on top with 4 counterweights!
At some point I decided to upgrade to EQ6pro, and I'm regretting it. It doesn't do well in RA, but is a very random problem that seems also depending on sky position. I've tried many things, maybe it's something I can't figure out but someone else can see. I do astrophotography since many years, but more pair of eyes are always better than just one.
You'll see the setup from attached pictures, it's free to rotate, no cables pulling/tense (I checked at night many times in addition to free rotation during the day), it's perfectly balanced in DEC and RA, and it has a total load of about 14/15kg (perfectly fine for a EQ6r).
I use PHD2 (pulseguide), both normal or dev with multistar, but I tried as well using The Sthe sky with relativeky X via ST4 direct, same problem happens. The problem is that RA guides always twice as bad as DEC. I tried taking PEC with Autopec in EQMOD several times, but it seems to not make it better, if anything seems worse maybe. PECprep said my PE was about 10/11 arcsec, so not too bad for an EQ6r from what I read.
The real problem remains: WEIRD peaks, and a general jumpy guide/graph, that make stars almost always slightly oval.
As you can see I check alignment with a Polemaster, and one night after the other it's basically almost perfect (Pole star almost central to the cross at the centre of the circle).
I'm in a Bortle 4/5 sky with good seeing, use a Lodestar X2 with 225mm guide scope, I used this setup for many years and never had a problem with my HEQ5.
The spikes seem worst when pointing near celestial equator, but also pointing elsewhere the stars are very slightly elongated.
I've done many other tests, but I'll start attaching some pics of my setup, and then a first test I did in a position of the sky more up (Bootes), and then something more down the celestial equator. For both I'll attach screenshot of PHDlog analyser and a 1:1 crop of the resulting pic with that precise guide.
I'm desperate, as it's not easy to get a replacement as these fly out as soon as they're sold. So if I can I'd rather solve it myself...but I don't know anymore I think I tried everything, and this is a some sort of substantial problem more than an optimisation needed.
Some pics of my setup:
And this is image/test number 1:
https://imgur.com/5rdEGDk (image crop 1:1)
https://imgur.com/y4WPU1r (sky position)
And finally this is image/test number 2:
https://imgur.com/HCoDrUM (guide...note those spikes!)
https://imgur.com/smfBmD0 (image crop 1:1)
https://imgur.com/OFZAWDH (sky position)