The Fury X Owners Club

[delshay]

After modding my Nano I made an error, but did not see it. To my surprised when I opened up HWMonitor, it showed Power, but no TDP readings.
I removed the fault at the controller "short circuit" & now HWMonitor does not show power symbol.

Question is why is power reading missing from GPU-Z & HWMonitor for Fiji chips.

New personal best not just in speed "1070MHz", but a new record low for temperature on air. It's going to get an even higher benchmark that's for sure.

Screenshot of benchmark with latest internal modifications.

 

[delshay]

New preliminary benchmark R9 Nano 1085MHz (305W+) needs big modification to hold this speed stable. I am working hard on a big update/fix..

Card can now run 1080MHz stable (300W). Like to see Fury x benchmarks.

 

[delshay]

My 300W+ R9 Nano is to mount a challenge to mighty Fury X including top score on HWBOT.

CPU increased to 3GHz compared to old screenshots.

 

[delshay]

WARNING: Do not attempt this on your R9 Nano as you need the correct equipment to carry out this modification.

R9 Nano is to get new capacitors. The old capacitor have all been removed from the PCB.

OLD CAPACITORS: All 3.5m ohms ESR or worst up-to 4m ohms with most showing 490uf.

NEW CAPACITORS: All 2.2m ohms ESR or lower with 520uf or higher (binned). To be installed on 18/06/2019.

.

 

[uuuaaaaaa]

WARNING: Do not attempt this on your R9 Nano as you need the correct equipment to carry out this modification.

R9 Nano is to get new capacitors. The old capacitor have all been removed from the PCB.

OLD CAPACITORS: All 3.5m ohms ESR or worst up-to 4m ohms with most showing 490uf.

NEW CAPACITORS: All 2.2m ohms ESR or lower with 520uf or higher (binned). To be installed on 18/06/2019.

View attachment 125211.

What did you gain with this mod?

 

[delshay]

What did you gain with this mod?

New capacitors have not yet been installed as I have more on order & I want to binn these to see if they are better than the ones I have already binned. They will be fitted tomorrow as this is when more capacitors will arrive.

The 16v capacitors there is a delay as they are direct from the AVX factory.
NOTE: The two extra capacitors on the rear of the card, no other R9 Nano has this. This has always been on my card.

 

[Fouquin]

NOTE: The two extra capacitors on the rear of the card, no other R9 Nano has this. This has always been on my card.

Those capacitors are on older production and sample cards previous to PCB Revision 10/11. The majority have the pads there with no capacitors mounted.

 

[delshay]

Those capacitors are on older production and sample cards previous to PCB Revision 10/11. The majority have the pads there with no capacitors mounted.

Those capacitors were added by me long time ago. They are going to be upgraded along with the high-side MOSFET capacitors to a new 220uf AVX with a tighter tolerance 10%. There are two other 16v capacitor internally which also will be changed. All 16v capacitors upgraded from the old 150uf. There's also new ceramic which should also be here tomorrow.

 

[Fouquin]

Those capacitors were added by me long time ago. They are going to be upgraded along with the high-side MOSFET capacitors to a new 220uf AVX with a tighter tolerance 10%. There are two other 16v capacitor internally which also will be changed. All 16v capacitors upgraded from the old 150uf. There's also new ceramic which also should be here tomorrow.

The previous revision test boards have them mounted, along with QC hardware, that's why the pads are present on those revisions. They were removed from later boards.

I assumed since you said that they have always been on your board that you had a QS card.

 

[delshay]

New re-binned capacitors installed, minimum value is now 529uf from 520uf, ESR unchanged. There's a little too much solder on two or three capacitors, but this can be cleaned up with ease. Rear capacitors removed along with all 16v capacitors on the front minus the fan capacitor.
Sorry for poor second photo, camera is just not having none of it (auto focus).

The very best binned capacitor is fitted to top (single capacitor), & the next best two binned (top right). The remaining 10 capacitors, the best two are sitting nearest to the GPU & the remaining 8 is random in the last 8 slots.

 

[delshay]

Here's an update on my R9 Nano with photo.

I got the wrong size capacitors so this card is running with just 2x 16v capacitors at the VRM, but it is supported by the other two on the back as seen in the photo. New & improved capacitors (correct size) on order, but will not be here until Jan 2020 next year. I was not able to fit 4x capacitors to the unused pads on the opposite side of the card, again I order the wrong size. Anyone that's interested in those pads they are 1.7v & 0.90v

Soldering quality of 16v capacitors does not matter much at this stage as they will be changed out again.

Cleaned up excessive solder with one or two more to do.

Card feels like it's more stable, but their is no change in undervolt & card is trying very hard at 1090Mhz overclock, but it still crashes, but the problem is pointing towards VRM running to hot. I have a fix for this, you may see an external heatpipe with radiator in the future.


SOLDERING/DESOLDERING:

To desolder those 3 contact point capacitors I had to increase preheater an extra 25c (375c). This is something i'v never done before as every desoldering I have done preheater is set to 350c. Having to much heat can damage the PCB as well as damage components, so be careful.

Also infrared lamp needed to be much, much closer to the PCB than normal, again this is not normal & can warp the PCB. For some reason the black PCB can tolerate such extreme heat. When desoldering/soldering be careful as you can warp or blister some PCBs even before the solder as liquefied, so take this one slowly.

Photo below.

 

[delshay]

R9 Nano now has a external heatpipe. This modification in it's current form looks really ugly as it did not know if it will work, so no photo. If I get enough request i will take a photo, but be prepared for nasty mod. Modification is having an effect on the VRMs, ie card is more stable at 1090MHz (max power), so we should see a new Superposition R9 record soon.

 

[uuuaaaaaa]

R9 Nano now has a external heatpipe. This modification in it's current form looks really ugly as it did not know if it will work, so no photo. If I get enough request i will take a photo, but be prepared for nasty mod. Modification is having an effect on the VRMs, ie card is more stable at 1090MHz (max power), so we should see a new Superposition R9 record soon.

I don't mind the nasty looks! Drop us a photo

 

[delshay]

I don't mind the nasty looks! Drop us a photo

Heatpipe modification has been taken out of service due this new modification for the core cooling. A new cooling solution has been found for the VRM. This is internal & can't been seen from the outside.

These new modifications are expected to bring improvements.

Ghetto Mods

That's awesome... aluminum radiators are so much more efficient at heat transfer (and cheaper to make) than copper ones I'm not sure why the PC insists on copper unlike virtually every industry (automotive, appliances, etc.) Because aluminum radiators are only more efficient in terms of size...

www.techpowerup.com

 

[gamefoo21]

Board number #000071

Even has the OG Cooler Master whine n dine pump... lol

It's my Furry, the modern version of my Rage Fury... Though I would have been so happy if the came out with Fury MAXX...

Flashed with the underclock, undervolt, and tightened timings BIOS. Living happily never going over 45'C

 

[delshay]

Board number #000071

Even has the OG Cooler Master whine n dine pump... lol

It's my Furry, the modern version of my Rage Fury... Though I would have been so happy if the came out with Fury MAXX...

Flashed with the underclock, undervolt, and tightened timings BIOS. Living happily never going over 45'C

That's a very good temperature. Can you show Superposition 1080p extreme with max power.

 

[gamefoo21]

That's a very good temperature. Can you show Superposition 1080p extreme with max power.

It'll be a bit, but those temps were with furmark and the power limit maxed out, it was reporting GPU usage at 97-100%. I even tried with the vulkan based test in the MSI burn app, and laughed the OGL was hotter by 2'C and slower.

Where do I grab the app?

 

[delshay]

It'll be a bit, but those temps were with furmark and the power limit maxed out, it was reporting GPU usage at 97-100%. I even tried with the vulkan based test in the MSI burn app, and laughed the OGL was hotter by 2'C and slower.

Where do I grab the app?

Download Superposition Benchmark from here https://benchmark.unigine.com/superposition

 

[delshay]

R9 Nano is currently getting it's final upgrade. "Correct" capacitors , power upgrade & maybe voltage tweak (undervolt in hardware). As i'm taking the card apart you will see some of it's secrets but not all.

Screenshot below reveals one of two internal cooling secrets which there are options.

My R9 Nano uses the "shroud" for extra cooling. The normal black pad has been removed & replaced with 17W m/k thermal pad 1.5 mm thick (stacked) x2 on the VRM heatsink.. The shroud is extremely hot when card is operating at max power/overclock. You will not be able to touch the card even after 3mins of power down. This shows the modification is working very well.

You can switch this to the GPU which shows clear cut 2c+ reduction when using LM, but for now do not put thermal pad on VRM heatsink & GPU heatsink together. ...Use one or the other. GPU requires two strips of thermal pad.

NOTE: Thermal pad must be place on the solid line (stitching part) of the heatsink but mine has stuck itself to the shroud. The stitching part of the heatsink is been looked at (CPU/GPU) on all heatsink, there's something i'm not happy about in this area on most if not all heatsink.

Anyway, if your GFX card has a metal shroud & it is not utilize very well into the design you could be in for a treat & a turnaround in your GPU temperature performance. This applies to most if not all GFX cards ever made that has a metal shroud.

 

[delshay]

ASSEMBLY

Card is ready for preliminary checks & is about to be put back together. VRM now has all four capacitors due to the old ones being to wide, which I was only able to fit two capacitors. One of the old wide capacitors has made its way to the fan headed. It's not very clear but the capacitor is push right-up against the connect, it's ultra tight as the plate needs to go on top of this.

New capacitors are the same as the old ones 16v, but with a reduction in size & ESR from 50 mOhms to 25 mOhms.

Upgraded power is also complete, it's just a matter can I control the heat. I will not know the answer until card is tested. Card is still awaiting undervolt in hardware & one more new additional cooling secret, bringing the total cooling secrets to three inside the card (not fully tested). which will be next & final, bringing this card modifications to a close.

 

[Peter Lindgren]

I did smaller mod to my Nano. Replaced the fan with a Noctua. The fan runs at 2000rpm which makes it silent. I have modified the bios and lowered the voltage and raised the TDP. The card runs at 980Mhz and never drops below that. The Noctua keeps it about 70c. Kind of ugly but it works very well.

I followed this guide but used a bigger fan => https://linustechtips.com/main/topic/690393-r9-nano-nuctua-fan-mod/

 

[delshay]

R9 NANO IMPORTANT UPDATE

If you are going to use the shroud as a heatsink, insure the fan cable is tucked down the side. This will give maximum contact & compression on the thermal pads.

GENERAL UPDATE (my card)

After doing the first capacitors changeover my card performed a little worse. After doing the second update with power increase it performed even worse, "throttling issues".

So today I conducted a full investigation why my card got worse. Investigation has found my card & probably all other R9 Nano cards to have a thermal pad issue. Test here has shown that R9 Nano cards should not thermal throttle with max power set in wattman until it hits the max temperature 83-84c (not overclocked). If your card is "overclocked", then it should not thermal throttle until it hits 76C+ regardless of how far you overclock the card with max power set in Wattman.

NOTE: I'm using 17W m/k thermal pads, but i'm also checking if FUJIPOLY pads are genuine, as it seems their maybe fake ones out there (not confirmed) awaiting test results.

Below is a dummy photo how I mounted the pads to fix the thermal throttling. The pads in the photo below is not perfect, but should give you an idea how to cut the pad to a perfect fit. The pads are fitted to the raised part on the internal metal plate.
Try not to exceed the circumference/perimeter of the raised bump. You can see some, but not all of the raised bump in the photo below, but there are six in total.

Your going to need a little bit of skill not to move the pads as you mount the PCB as you can not compress the thermal pads manually when using 17W m/k type of thermal pads.

Photo includes one my pride & joy tools, Gerber 650 Evolution (ultra rare) full kit.

 

[delshay]

Just a small update, list of parts I fitted to the card.

EEF-GX0D561L Panasonic | Mouser

EEF-GX0D561L Panasonic Aluminium Organic Polymer Capacitors 560uF 2.0volt datasheet, inventory & pricing.

www.mouser.co.uk

& here. Note the small numbers in stock. That was me, I order them back in July https://www.mouser.co.uk/ProductDetail/AVX/TBME227K016CBSZ0000?qs=sGAEpiMZZMsh%2B1woXyUXj1w%2BO54z6D%2BmAyhb0Qy3w0A=

3DMARK stability open test bench (max power) not overclocked.

 

[Valantar]

Anyone got any tips on undervolting my Fury X? I've tried before (IIRC with both Afterburner and through the driver software, and while both seemed stable in testing (around 1.05V IIRC), they weren't in games (crashes and instability in Rocket League at least, probably some other titles I can't remember too). I think I remember someone posting on these forums about an UV BIOS and having success with that, but for the life of me I can't find it. Any help would be much appreciated.

 

[Kanan]

Anyone got any tips on undervolting my Fury X? I've tried before (IIRC with both Afterburner and through the driver software, and while both seemed stable in testing (around 1.05V IIRC), they weren't in games (crashes and instability in Rocket League at least, probably some other titles I can't remember too). I think I remember someone posting on these forums about an UV BIOS and having success with that, but for the life of me I can't find it. Any help would be much appreciated.

Undervolt is undervolt, whether with software or by firmware, doesn't make much difference. I would simply increase the voltage until it is stable again.