Have 200 MBPS service, but brand new Arris SB8200 modem tops out under 100 MBPS downstream

I just bought an Arris SB8200 modem (+ Tenda AC5 v3 router) to replace our rented Arris TG1682G xFi gateway.  We have 200 MBPS service, and the TG1682G reported downstream speeds in the 150-200 MBPS range.  The SB8200 is consistently reporting speeds in the 70-100 MBPS range.  

I'm not seeing any lapses in Internet connectivity or other instability, at least externally.  (See event log summary below.)

I'm wondering whether there might have been a provisioning error -- here's why:

We live in a part of town with wretchedly poor cellular coverage.  To deal with that, we run an Ethernet cellular signal extender (a Samsung SLS-BU10B, purchased from our cellular provider, Verizon) that gives us 5 bars of cell signal within our home -- but requires an Internet connection to do so.  "Landline" phone (mainly for occasional faxing) is VoIP, also dependent on the Internet connection.

So when I called Comcast Customer Solutions to get the new SB8200 modem set up, what the agent did for me -- so that in case of failure we would not be left without a means of contacting Comcast -- was to temporarily set up the SB8200 as a second device on my account, so that I could get that running, and then de-provision the rented TG1682G gateway.  I got the SB8200, the Tenda AC5 v3 router, and all the devices on our home network running, but I noticed that the SB8200 was testing consistently at just under 100 MBPS.  I assumed that it had been provisioned as a (temporary) second device at a default 100 MBPS.  When I called back to get the rented gateway de-provisioned, I asked that the SB8200 be provisioned for 200 MBPS, since that's what we're paying for.  The agent told me that the speed increase should happen overnight.

Next morning I discovered that Comcast had pushed a firmware update to my SB8200.  Out of the box, my SB8200 did not have a login screen on its web interface; now, it did.  Arris implemented the login screen in a firmware update released apparently after my SB8200 was manufactured.

But no speed increase. 

This could be that it hasn't been provisioned correctly, but I'd like to rule out anything that might be wrong with the SB8200 unit itself or with the RF signal it's seeing.  So here's a summary of channel status and event log entries:

I see 32 locked QAM256 downstream channels, ranging in power from -7.0 dBmV to -2.8 dBmV, with SNR ranging from 38.5 to 41.2 dB, ranging from 0 to 5 successful corrections per channel and 0 uncorrectables over the course of 6 hours of observation.  There's also a single DOCSIS 3.1 OFDM channel reporting -5.1 dBmV power, 36.0 dB SNR, and a ridiculously high successful corrections number which I assume is meaningless garbage.

There are 4 locked QAM256 upstream channels, ranging in power from 40.0 dBmV to 41.0 dBmV.  According to Arris tech support this is low, but it's within Comcast specs.

The event log has the following types of events (and number of occurrences):

"SYNC Timing Synchronization failure - Failed to acquire QAM/QPSK symbol timing" - 15 times over 4 days.  Timestamp for this event doesn't ever seem to be set -- always displays as 1970-01-01 00:00.

"No Ranging Response received - T3 time-out" - 12 times over 4 days.  Also no timestamp set.

"SYNC Timing Synchronization failure - Loss of Sync" - 4 times over 4 days.

"TLV-11 - Illegal Set operation failed" - 4 times over 4 days.  Always followed by...

"Config File Rejected due to Invalid or Unexpected TLV 11 encoding" (- 4 times over 4 days).

"Honoring MDD; IP provisioning mode = IPv6" - once, during troubleshooting the apparent 100 MBPS max speed with a Comcast tech support rep, who ultimately said she'd "run to the end of her script" and that I needed to pursue the problem with Arris tech support :-P.

"REG-RSP-MP Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value" - once, during that same troubleshooting session

"DBC-REQ Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value" - once, during that same troubleshooting session

Does any of this event log activity suggest either (a.) a provisioning error, (b.) a deficiency in the RF signal that is causing the modem to max out at < 100 MBPS, or (c.) a defect in the modem itself?

One final thing: I'm running the speed test from a Windows 10 computer plugged into a LAN port on the router.  I tried plugging the computer directly into the cable modem but couldn't get an IPv4 address assigned via DHCP (despite the router being able to do so without fail, when *it* is plugged into the cable modem).  Running ipconfig in an administrator cmd.exe shell when connected directly to the modem, I see a link-local IPv6 address and an fe80:: prefixed IPv6 default gateway address, but nothing else.  Any idea why I'm not getting a DHCP address?  (I disabled APIPA IPv4 169.254.x.x autoconfiguration, which was kicking in before any DHCP response when I tried to connect the computer directly to the modem; do I need to disable IPv6 autoconfiguration too?  If so, how do I do that?)

Thanks for any help that can be provided!