I am going to make your choice so much easier right now

I just bought RME DSPE Raydat, and I will tell you my view on it, this might actually complicate your question, but it might also clarify A LOT of stuff. With my personal experience.

You are doing a side-grade, not upgrade or downgrade.

I am currently 3 days in after purchasing RME DSPe Raydat

I needed to use my old Saffire Pro 40 as a speaker driver (as a speaker amp essentially) with ability to use pre-amps and route the signal back to RME DSPe Raydat PCIE card. I really like Saffire Pro 40 for it's bass response and power of midrange. Albeit highs are subdued, which I like for my ears. Gives the sound very tape like saturation quality with my Yamaha MSP5s (arguably better for listening than most newer Yamaha speakers like HS5 or HS7), but I digress...

I also had bad clocking issues, because as most know that Focusrite Saffire Pro 40 is Firewire, so I was like:

"Well, let me clock to PCIE card, and also not get rid of Saffire Pro 40", I refused to buy Babyface, because it looked like spaghetti monster with side mounted cables. Not my cup of tea.

I tried 2 variants of my set up:

1. I ran ADAT out and ADAT in from RME to Saffire Pro 40, that was then master clocked to RME Raydat. At first I couldn't have a proper clock and I kept getting pops and clicks. But due to some hardware updates I had to reinstall Windows 11 and issue disappeared. As I could not clock to Internal clock of RME, while feeding ADATs into Saffire. Now I can.
That result does alter a sound to a noticeable level. Lows are not as deep, midrange is not as full (as Saffire acts as a speaker driver now), but overall sound is more clear on all spectrums.

They say RME does not color the sound, but I disagree, it has to go through RME's converters and chips still, so it does change sound, I clearly hear it. I am comparing Pepsi to Coke here, and I can definitely tell how Pepsi tastes like, and I know how Coke tastes. I honestly prefer Pepsi

What is Pepsi here? Ah, who knows!

a) Because when I listen to low end frequencies and want the best midrange fullness, I get the best sound design with plainly plugging Saffire Pro 40 through Firewire into a Firewire PCIe card.
b) When I want clean sound, RME HDSPe DEFINITELY takes the cake, but I hear transients are sharper with it, and midrange is not as rich, lows are not as bassy and lush.

In my case you have another stage that is affecting sound: DA converters of the Saffire Pro 40, they change the sound before it hits the speakers, running ADAT in from RME to Saffire, and sound now coming out of Main Outs of Saffire.

Before someone says that cannot be, it can be, as even running SPDIF digital output from SSL Alpha Channel preamp to digital input makes the sound clean but very cold, compared to running Analogue of SSL Channel into Saffire and having it convert there and come out through Firewire into the Cubase 5.

I like the lush sounds of Saffire AD conversion, but it has those low end and midrange qualities.

2. Running DSPe by itself as a master, and having Saffire Pro 40 as secondary device makes the sound clean and very good, but arguably not better than Saffire 40, though muddiness is gone when using DSPe.

3. Third way I was able to clock my Focusrite Saffire Pro 40 with almost no drop outs was to have my Gigabyte Aorus Master Z690 motherboard Optical SPDIF run into ADAT of Saffire Pro 40, then clocking to Optical SPDIF with Saffire Pro 40 to motherboard. This motherboard in particular has ESS SABRE Hi-Fi 9118 DAC chip, which is used on many top end interface units in Pro Audio, according to research.

So, in my case it is the closest sound I can get to my Saffire Pro 40 with zero drop outs (but I must keep firewire cable plugged in while Master is my Optical SPDIF of the motherboard), but I ran HDSPe with Saffire 40 to compare.

It's 3 types of flavors

Neither wins in everything.

Here is my answer in Layman's terms (Pros and Cons)

RME HDSPe
Pros (by experience):
1. Incredible latency: 0.688 ms Input, 1.365 ms Output, in Cubase 5, 96khz 24 bit. Hard to beat by anyone.
2. Clean sound, tight transients, albeit a bit cold in comparison to some units. Not muddy.
3. Powerful clock
4. Low jitter
5. TotalMix routing
6. ADAT for days! To connect ADAT gear. You can also run ADAT out to DAC units and get your sound even better than most audio interfaces.
7. Expansion card that comes with it, for another ADAT expansion.
8. Stable drivers
9. Way more inputs than Babyface

Cons:
1. You don't have pre-amps
2. Set up may present some issues if you have some driver issue in the system or have not updated firmware. Make sure you do this first.
3. NO EQ in TotalMix
4. No TRS line in
5. No speaker knob to control volume, must have a speaker amp, another interface such as mine, or mixer to drive speaker volume (if you like analogue volume adjustments)

Baby Face 
Pros (by theory)
1. EQ available in TotalMix, can hone your speaker to the sound you like
2. Stable drivers.
3. 2 Pre-amps
4. Screen with speaker volume dial.

Cons:
1. Less Inputs/Outputs than HDSPe
2. Cannot compete with HDSPe's latency, because it is USB
3. Spaghetti wires out of the sides of the unit, don't like that at all. Some don't mind.
The box is small, so connectivity from the back only was not possible

I don't like Babyface for the way wires are routed. It will be a mess on my desk.

You need to watch this review on Babyface:
https://www.youtube.com/watch?v=3i5iS1ULI6Y

When I typed out this narrative, it has also made me trully consider HDSPe AIO Pro as a swap option between Raydat and UCX II that I am considering.

Long answer, but I doubt anyone will type out so much to tell you about this.
It's so hard to get answers on forums these days.

This is my 2nd post on the forum., and I hope that helps you out :-)

Thanks for the replies
As I said I already own the babyface and I just conider swapping it for the old Digiface HDSPe with the cardbus express (and that going to a simple cardbus to pci-e adaptor for connection to a pc)
Ok the fx and the eq of the total mix were a really useful asset but i could live with it. Internal preamps is something i don't care any more since i use external preamps etc etc.

So in my eyes the only issue would be if the old DIGIFACE is equally futureproof, stable and reliable as Babyface.

Thanks for the input. Babyface has to go since I need more in and outs. The other option I have would be a digiface usb, but then I go for 450 euros  (since there is no swap offer and babyface sale is not guaranteed) and there are very few in the used market at the moment since it's quite new  (so if anything fails this will be my future investment, since it will be more common in the used market in 3-5 years time).

Thanks, I edited my post, so not to mislead anyone.
I was convinced HDSPe AIO Pro had EQ.

Thanks for the input but raydat is not for me because:
It has the same  amount of Adat in outs. I don’t need any more midi in and outs, since i have dedicated interfaces for that. No need for aes , spdif since i work mostly analog gear.
Why should i need a word clock, other than the fantastic clock of the rme card (the converters are always slaved to the rme interface)? So it seems like an overkill. That’s why i still consider the older gear and for new only Digiface usb is the one.

I personally believe that in theory digital to digital should be the same, but yet SPDIF Optical out of motherboard will sound very different compared to ADAT out of HDSPe Raydat into Saffire Pro 40 (as standalone unit), I listened to both. I had first scenario as Motherboard Master using SPDIF Optical of the motherboard, and on 2nd scenario I had Raydat as Master.
Motherboard uses ESS SABRE Hi-Fi 9118 DAC chip, which is a great chip from what I hear. I can't speak on differences between the chips though.

There is also electrical resistance differences, they will impact phases and other THD differences.

Raydat produced cleaner signal (removed a lot of muddiness of Saffire Pro 40), but higher noise floor than Saffire Pro 40. It was like I am using a completely different Interface and bypassing Saffire.

Raydat altered frequencies of my Saffire Pro 40 being in standalone mode. I do believe whatever device you are hitting from Raydat or Motherboard plays a big role in how source clock "sounds" as well, as sync can be not exact, and creates a phase somewhere in the signal. Saffire Pro 40 had issues with maintaining lock while using Raydat, and would create metallic noise, until I reinstalled Win 11.

In my case: motherboard Master produced closer sound to original Saffire Pro 40 if I was coming out Firewire end into PC.
RME produced a much cleaner sound, with better transients, but worse low end and way thinner midrange.

Digital to digital is not the same, and there could be differences when syncing over SPDIFs too, which I am going to test today actually.

There is a big difference between ADAT sync and straight in using Firewire in my case.

I have several records I know like the back of my hand and I kept testing it. Syncing to ADAT vs Internal on RME also produced slightly different acoustical results, with Internal being better in terms of fullness, but produced thinner but cleaner sound with external sync to Saffire Pro 40.

Digital to Digital is not the same, in theory it should be, but it's not.
Firewire does not = ADAT
ADAT does not = SPDIF Optical
USB can sound different than Firewire.

Protocols make the difference too, it was an eye opener to me with my Raydat.

Also to take into consideration, is your DA conversion skipping mosfets of the unit somewhere in the signal chain?
According to Focusrite it does not, but I call it bluff on the aspect of not knowing their own 15 year old unit.

To me Saffire Pro 40 sounds better than Scarlet, and that's an arguable point, but Scarlet is hardly an improvement to me, over Saffire Pro.

I guarantee you that ADAT out and in from Scarlet will sound completely different than Saffire Pro 40

I think the answer is very obvious, be that admin or not, as PCI is way older principle and cannot maintain same level of speed/transfer

32-bit PCI speed is 133 MB/s, while the 64-bit PCI speed is 266 MB/s

PCIE on the other hand

Most PCIE devices for audio are running at 1x PCIE Gen 3
You will begin to notice pops and clicks and slow downs as you hit more channels than 266 MB/s can handle.

It is definitely possible

You live your life, by all means.

To me it's very clear, I hear difference of 0.2db on instrument gain changes, and that is also a "myth", and "cannot" be heard.

According to some audio engineers all 0s and 1s are identical, yet, no one truly understands it on electrical level.
I don't mean to insult you either, but if you believe what you believe, you are allowed to.

But...many people can't even hear difference between Scarlet and Clarett by Focusrite, so why would people hear even more nuances of going from different forms of AD and DA conversions?
They won't, cause they can't.

I am not even considering myself at the top of the line of engineers.
A friend of mine hears at which frequencies you have phases of instruments, and can EQ it out with ease.
He knows where it is, and can clean it up.

Digital signal is more complicated than 0 and 1, and way more complex than just the converter.
There is a reason why many headphone amps treat headphones differently, based on their resistance.

Analogue has characteristics, and electrical resistance makes a difference.

Clocking Saffire to RME through ADAT changes the sound, even though I am not sending a signal to RME. Sending clocking signal from RME to Saffire does change the sound of Focusrite Saffire Pro 40 main out

If I didn't know my equipment for 15 years, I wouldn't speak on it.
I tried about 5 scenarios with routing, and they all produced slightly different results in audio from main out.

I think audio engineers are one of the most stubborn people I have ever met, because only few are open minded to explore electrical engineering behind their own equipment. That rabbit hole is deep, and there is a reason why many pros got their way of routing channels. There is nothing crazy with it.

I used to overload mixes to get harmonic distortions, and was told not to do it, and being told that I am not bright for it.
I liked the sound behind certain voices for it. 2 days ago I saw Dr. Dre talk about how he would overload mixes, to get his drums punchy.

If I can hear the difference, that's all it matters. How you spend your money is up to you.
I know a difference between digital signal and analog, but I also understand that every equipment is unique.

For an example my Alpha Channel has completely different sound if you go Post or Pre-EQ, and whether you bypass it or not, just putting signal through post EQ will alter sound, but EQ can be completely off.

Magic? Nope. Signal chain, and how the unit is built.
Imagine thinking that all digital signal is the same.

Amazes me how people are not willing to try their own equipment to tell the difference.
Me on the other hand, I understood how simply hitting SUM on SSL Alpha Channel gave more grittiness, and gave more harmonic distortions.

How I discovered it? By trying your own equipment, quit making people that know their own equipment sound like crazy people.
That's uninspiring at best

Most can't hear 48khz vs 44.1khz, but doesn't mean everyone can't. You going to call those people crazy for hearing the difference?

AD/DA conversion and ability to sync produces much more defined differences in sound. Even syncing one device to the other can alter the sound response. 

Most recording engineers are not engineers truly, their mind is not for engineering. It's for creating music. The science behind sound devices is vast. There is a reason why electrical engineers build these units, and audio engineers just come in and say if they like them.

My hats off to those that design these units and can tell a difference.
After being in the industry for a while, and seeing people design headphones as actual engineers, modding microphones and such, I don't demean a person for hearing something that I cannot.

But...most can barely hear a difference between Yamaha NS10 and Yamaha HS7. But, what never ceases to amaze me about the industry is how people are quick to judge someone who can hear a plain difference in sound.

I circle a lot of forums from Gear Space, to AudioScienceReview, and last thing I would say to a person is that they a full of it, for hearing something that I could not.

He is my long answer to you, about your so called "misinformation".

I urge everyone to know their gear, be that a tube amp that you love to overload, or a piece of digital gear like DAC or RME unit

And I just meant the in regards to the op question, wich digiface to get it was complicating matters to much. Like saying proper grounding and power distribution has an influence too. Yes it does but not the question here.