Comparison with an OEM pad
Generally speaking, the PTM7950 is currently only available with a thickness of 0.25 mm, everything else is not original. However, you can also find offers with 0.2 mm on the usual suspects (AliExpress, eBay, Amazon, etc.). These are usually also PTM pads, but without a clear origin and performance. So you can safely keep your hands off all pads with 0.2 mm if you want the original. As luck would have it, I also have PCM5000 and PCM8500 pads in my drawer of good deeds for my upcoming roundup and I noticed that one of the Ali pads sent to me in PTM7950 guise had various similarities with the PCM8500.
Visually, the colors are not that far apart and if you don’t have the PTM7950 for a direct comparison, you won’t notice any real difference. That’s exactly why I took a microscope picture of the two pads. On the left we see the PTM7950, which has a slightly different grind and apparently a different matrix. You can also tell that it is a different polymer by its consistency. Although the PTM7950 is thicker, it is significantly softer and easier to bend. It is also stickier and tears more quickly, whereas the OEM pad is more likely to break.
So let’s take a look at what’s inside. The fillers are relatively similar in proportion, but unfortunately I can’t examine the polymer with my own resources. But even so you can see small but subtle differences, let’s start with the Honeywell PTM7950:
The OEM pad has a slightly higher proportion of polymer, and there is also a slightly higher proportion of zinc oxide in the fillers. The rest looks about the same.
The measurement brings it to light!
The long burn-in cycle at 8.5 Newton and from 25 to 75 °C, on the other hand, is clear. We can first compare the thermal resistance (yellow and red curve) on the left Y-axis. Then add the dotted lines (blue and white) of the layer thickness (BLT) for the right Y-axis:
The whole thing can then be summarized as follows, whereby you can see that the PTM7950 has a significantly larger temperature window for the phase change and at the end also the thinner layer thickness (BLT) and a slightly lower thermal resistance:
Honeywell PTM7950 | Noname OEM Brand | |
BLT without pressure |
250 µm | 200 µm |
BLT at 25 °C (8.5 N) | 106 µm | 153 µm |
Start burn-in | 40 °C, 38 µm BLT | 40 °C, 45 µm BLT |
Burn-in | 45 °C, 18 µm BLT | 43 °C, 36 µm BLT |
End of burn-in | 50 °C, 16 µm BLT | 45 °C, 36 µm BLT |
Final thermal resistance | 0.08961 K/W | 0.09494 K/W |
Summary and conclusion
The PTM7950 (theoretically) has the ability to shrink to a thickness of up to 16 µm at less than 10 N per cm², i.e. the values that we also find on the graphics card, which is excellent and is in line with the value of more liquid pastes. However, you always have to take into account that the surfaces on the CPU and GPU are not straight and flat, but often curved. The PTM7950 is still quite flexible and should also be able to compensate for certain material expansions of the cooler base. However, the OEM pad shown here is as hard as a board and tears quite quickly.
Where and whether you use these pads is up to you to decide. With a smooth radiator base, good contact pressure and a rather unbent GPU or CPU, the pad can certainly be used well on all types of water cooling systems, which nobody has denied, by the way. However, I also wrote that I wouldn’t do it with the custom loop, but would rather continue to use a viscous paste. A matter of taste. The reasons lie not only in the burn-in, but also in the unpredictable change in layer thickness. 250 µm can quickly become 16 µm and then you have an uncalculated height change in the radiator structure (stopper!). Or you may have got a pad from the Ali shelf that is too brittle and have other problems.
The distance between the heatsink and memory modules also decreases significantly with GPUs and you may have to adjust these pads on the memory in advance so as not to put too much pressure on the memory. Or these pads are too hard and you don’t get enough pressure on the phase change pad. What works on the CPU and an AiO can, with a bit of bad luck, also lead to problems on large water blocks. Once you have filled a custom loop, things get complicated. I just want to point this out, but I don’t want to stand in the way of anyone’s desire to tinker. It will certainly work in most cases, but it doesn’t have to.
As for the replacement pad: it’s not that bad and if the price is right, it would work too. Especially as it would be much easier to apply. But then you have to market it honestly as a PCM8500 and at a lower price (if I’m right) and not as a Honeywell PTM7950. In the end, that really is cheating the customer.
4 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Neuling
Urgestein
Veteran
Alle Kommentare lesen unter igor´sLAB Community →